EAAP36 Conference Programme

Abstract:
Fear of flying (aviophobia) is a common and often disabling anxiety disorder that can significantly restrict personal, social, and professional functioning. Despite its high prevalence, many clinicians feel insufficiently equipped to assess and treat this specific phobia effectively. At the same time, aviation professionals frequently encounter passengers experiencing significant flight-related anxiety, yet may feel uncertain about how best to respond.

This workshop presents an evidence-based and clinically applicable approach to the assessment and treatment of fear of flying, drawing on cognitive-behavioral therapy (CBT), exposure-based interventions, and trauma-informed care. The VALK Foundation brings over 36 years of specialized clinical experience in the successful treatment of fear of flying and associated underlying psychological conditions.

Participants will gain insight into the multifactorial nature of aviophobia, including fear of loss of control, panic sensations, catastrophic misinterpretations of bodily symptoms, and trauma-related factors. Special attention will be paid to differential diagnosis, comorbidity (e.g., panic disorder, PTSD), and medical considerations that may complicate treatment.

The workshop will cover:

Comprehensive assessment and case formulation of fear of flying
Psychoeducation tailored to aviation-related fears
Cognitive interventions targeting catastrophic beliefs
Graduated and in-vivo exposure, including flight simulation and real-flight exposure
Managing setbacks, safety behaviors, and relapse prevention
Practical guidance for aviation professionals: how to recognize severe flight anxiety, effective communication strategies, and do's and don'ts when supporting anxious passengers
Through clinical examples and interactive components, participants will learn how to translate theory into practice and adapt interventions to individual needs. The workshop is relevant for clinicians working with anxiety disorders as well as aviation professionals who wish to better understand and support passengers with fear of flying in an effective and ethically responsible manner.

This workshop will introduce key considerations for integrating Human Factors (HF) into engineering processes for flight deck design and certification and summarize essential concepts. This includes choosing the right approach as well as addressing essential HF aspects.

Summary of aims:

Achieve a global understanding of the connection between engineering, certification and human factors with practical examples

A wise person once said "the only people who actually like change are babies with a dirty nappy!" And it's true, we humans are not all that fond of change. Just think about it, in general, and certainly if we are satisfied with a situation, product or way of working we like things to stay how they are.
Change is often met with a great deal of resistance by those that face it. However, if a change is relatively simple and we actually want it, let's say for instance changing to a new car, then you can often just get away with putting the change in place.
However, if the change is in any way complex, or if the willingness to change is less than high (or a combination of both) then change can be difficult to implement. BUT with a little TLC this can easily be helped. The trouble often is that in technical, procedure driven, safety focused organisations, giving their humans-in-the-loop the TLC they need, is often overlooked.
So, what CAN you do to make sure employees adopt, accept and even embrace a new way of reporting, or that new VR trainer, an organisational change or any new way of carrying out their work?
In this interactive workshop NLR introduces their simple stepwise methodology with tips and tricks to reduce resistance to change and help with the adoption and acceptance. Based on academic change models but put in to a workable model that's even aviation themed. It focusses on involving your stakeholders, creating understanding for the urgency of the change and reducing resistance. Making change a lot easier on everyone!

The aim of the workshop will be the exchange of knowledge and experience with respect to the future of aviation in the cockpit between a variety of specialists. Topics:
• Increased use of Digital Assistants
• Human AI teaming in practice (cockpit, TWR, ACC and airport)
• Reduced Crew Operations & SPO

We propose to first briefly present four different SESAR and Horizon Europe-funded projects, together with some additional information about modern developments in automation, which all have the three above mentioned components in them. Those projects are:
• JARVIS (Just a rather very intelligent system)
• RESPONSE (REduced or Single Pilot Operation iNcapacitation Safety Enhancement)
• HAIKU (Human AI teaming Knowledge and Understanding for aviation)
• HUCAN (Holistic Unified Certification Approach for Novel systems based on advanced automation)

After a brief introduction of those projects we would like to have a number of break-out moderated discussions on subtopics with the audience. The amount of participants would influence how many subtopics we will eventually use. The idea is that there will be the option to choose the topic of preference and also to move from one topic to the other.

For these subtopic discussions we would like to use flip-charts with moderators to encourage the audience to exchange opinions. The subtopics will be:
• What do pilots, air traffic controllers and airports really need from automation and AI?
• What are the top ten key aviation Human Factors research questions for the coming two decades?
• What is the best human-AI teaming relationship?

Afterwards the moderators will feed the results from the discussions back to the whole group of participants, and there can then be a panel discussion (panel with audience Q&A).

Eventually all participants will have a better insight in the automation / AI related future of aviation so that they can use that for their own roadmaps and decisions to make about next steps in their research.

The target audience for the workshop will be EAAP36 conference participants who are relatively new to the aviation industry and/or to aviation psychology and human factors and are seeking an introduction and overview of current topics in aviation psychology and aviation human factors as related to safety.

Workshop Summary:
Proposed topics include:
• Introduction to aviation human factors;
• Human factors and safety events;
• Human factors training applications;
• Systemic approaches to safety;
• The positive contribution of humans to safety.

Summary of aims:
• Provide an introductory overview to aviation human factors.
• To develop participant knowledge and provoke thoughtful discussion on current topics.

Workshop led by:
Brenton Hayward & Dr Alan Hobbs

Overview of how to construct and use a task analysis to develop a pilot selection system. This includes finding resources, identifying appropriate taxonomies, developing the preliminary task analysis, selecting participants for the evaluation group, scoring the results, identifying the relevant KSAOs, and constructing a preliminary list of selection instruments.

At the end of the class, the participant should

- Understand the importance of a task analysis for developing a pilot selection system
- Know where to find supporting documentation
- Be able to construct a preliminary task analysis
- Select appropriate SMEs and score the results of the preliminary task analyses
- Identify the relevant KSAOs
- Construct a preliminary list of selection instruments

The working group was commissioned by the board of DGLP (German Society for Aviation Psychology) to develop and submit a curriculum and introductory courses as part of a training and professional development program for Operational Aviation Psychologists (OAP) and Clinical Aviation Psychologists (CAP) based on the adopted accreditation scheme developed earlier.
The content of these developments has been coordinated with the E.A.A.P. from the outset. Our sister organisation, the AAPA (Austrian Aviation Psychologists Association), is involved in the course development and implementation. Our goal is to work towards a shared understanding of the essential elements and content of aviation psychology within the broader community and to put this understanding into practice. The collaboration with E.A.A.P., with national professional societies and organizations such as the German Federal Armed Forces (Air Force), and academic institutions are part of the program.
The courses are designed to provide an introduction to the fundamentals of the following important areas of Aviation Psychology. (1) Overview - Aviation Psychology; (2) Aviation Knowledge Basics; (3) Human Factors and Cognitive Processes in the Work Environment of Aviation Professionals; (4) Psychological Assessment of Operational Staff in Aviation; (5) Crew / Team Resource Management, Training & Communication in Aviation; (6) Aviation Safety and Risk Management; (7) Clinical Psychological Assessment and Intervention in Aviation.
In line with our members' specific needs advanced studies of individual methods, such as CISM or CRM/TRM, will continue to be offered in separate follow-up courses. The DGLP course program is intended to culminate in an examination for accreditation as an OAP or as a CAP.
Members of the working group will present the current status of this project together with an implementation program of practical courses for psychologists who wish to become certified aviation psychologists. The curriculum will be discussed with the conference participants for broader exchange and receiving feedback, which will be taken on board in the further practical implementation and further collaboration and developments for a common way forward.

Aviation mental health has received increasing attention as researchers and regulators examine how psychological well-being may influence aviation safety. High-profile aviation events highlight mental health symptoms in aviation incidents, while recent research has identified healthcare avoidance and barriers to care among aviation personnel. However, accident investigators currently lack a standardized method for assessing whether mental health factors may have contributed to aviation accidents. Investigators also report hesitancy in addressing these issues due to limited training in mental health and human performance. These challenges may contribute to inconsistent documentation and limit data needed to identify safety trends. To address this gap, an international multidisciplinary research team developed the Accident Investigation Mental Health Job Aid (AIMHJA).

METHODS

The AIMHJA was developed using a six-round modified Delphi process to establish international expert consensus among specialists from multiple aviation domains. Participants included aviation psychologists, mental health practitioners, accident investigators, regulators, pilots, and industry representatives. The modified Delphi included four online mixed-method questionnaires and two in-person consensus sessions conducted at international aviation conferences. Across the questionnaires, 127 validated responses were received, with 53 subject-matter experts participating in facilitated discussions. Participants iteratively developed, revised, and evaluated AIMHJA prompts and guidance using Likert-type and open-ended feedback.

RESULTS

The consensus process resulted in a two-tier investigative job aid. The first tier provides screening prompts for investigators without formal mental health training to help determine whether consultation with a specialist is warranted. The second tier provides specialists with a structured set of prompts for follow-up interviews. Expert recommendations focused on refining question wording, incorporating culturally appropriate language, clarifying investigative scope, and formatting the tool for integration into existing accident investigation workflows.

DISCUSSION

The AIMHJA provides investigators with a structured framework for considering mental health and aviation psychology factors during accident investigations. The framework may support investigators in addressing sensitive mental health topics and promote more consistent documentation of human performance factors. Standardized documentation may support identifying broader trends and inform future aviation safety recommendations. Future research will focus on validation and implementation studies evaluating the practical application of the AIMHJA.

Background: Effective pilot mental health management must be proactive, operationally feasible, and non-stigmatizing. This study describes a psychological risk assessment and intervention system developed through long-term operational use across multiple airlines. The system is designed to enhance safety while protecting pilots' careers and minimizing operational disruption.

Methods: The program was implemented across three airlines over 6 years and included more than 7,100 pilots. The system operates through a three-level, dimension-based, closed-loop framework designed to ensure that pilots receive structured, supportive, and recovery-oriented care at each stage of the process. The first level uses digital screening to establish individualized psychological profiles ("one-person-one-file"), enabling longitudinal tracking. This level delivers personalized psychological health reports via a 24/7 accessible digital platform and incorporates rotating assessment scales to enhance monitoring sensitivity and reduce assessment fatigue. In addition to individual feedback, aggregated organizational insights are generated for preventive management purposes. The second level conducts multi-method, group-based supportive re-evaluation to clarify initial findings and reduce over-alerting. The third level provides structured, one-on-one supportive re-evaluation for higher-concern cases and facilitates referral to specialized care, with aeromedical coordination when appropriate. Across the system, cases are managed through standardized assessment, follow-up, and reassessment procedures, including return-to-duty monitoring after medical clearance when applicable.

Results: Across repeated assessment cycles, pilots showed increased willingness to disclose personal psychological concerns. Indicators of psychological health showed measurable improvement over time, while broader well-being metrics remained stable within a healthy range. The closed-loop workflow improved the consistency and practical utility of case management, strengthened confidentiality and data governance, and addressed key implementation challenges, including over-alerting, resource constraints, anxiety amplification, and concerns about career impact.

Conclusion: This three-level, dimension-based, closed-loop framework operationalizes pilot psychological assessment as a structured, governance-ready risk management process. It provides a practical model for airlines seeking proactive, human-centered psychological safety risk management aligned with operational realities.

In the civil aviation sector, pilots' stress management and coping strategies are critical components of effective risk control. To better understand civil pilots' stress management and coping processes, their self-care ability may serve as a sensitive indicator of psychological well-being. In a previous study, we found that civil pilots' self-care quality not only reflected their levels of self-awareness and self-disclosure, but was also positively associated with resilience and negatively associated with fatigue.
In a super-aging society such as Taiwan, promoting successful aging has become increasingly important. The selection, optimization, and compensation (SOC) model proposed by Baltes and Baltes (1990) has been widely validated as a reliable framework for assessing successful aging. Therefore, it is necessary to examine how civil pilots' self-care relates to their SOC scores. Furthermore, it remains unclear how civil pilots' self-care ability compares with that of other professionals.
To address these questions, we administered a 45-item self-developed questionnaire that included measures of self-care (i.e., self-awareness and self-disclosure), SOC, and sleep quality. The questionnaire was designed to assess participants' self-reported self-care ability, successful aging, and health-related indices. A total of 522 participants (287 civil pilots and 235 non-pilot professionals), aged 22 to 65, provided informed consent and completed the survey.
We hypothesized that civil pilots would demonstrate higher levels of self-care, sleep quality, and successful aging compared to age-matched non-pilot professionals. We also hypothesized that self-care ability would predict sleep quality and successful aging outcomes, regardless of participants' occupation or chronological age.
The results indicated that civil pilots scored significantly higher than non-pilot professionals on both self-awareness and self-disclosure. Similarly, civil pilots reported higher SOC and sleep quality scores. Multiple regression analyses further revealed that SOC and sleep quality scores increased as a function of self-care among civil pilots. Implications for incorporating self-care consultation into pilot support programs to enhance civil pilots' mental and physical health are discussed.

This paper looks at the prevalence and characteristics of sim anxiety - the sometimes severe anxiety experienced by pilots when they undergo their regular OPC or LPC simulator checks. Anecdotal evidence has indicated that there is high prevalence of this amongst pilot at all levels of experience at varying levels of severity. The survey will look at the symptoms experience, the antecedent events that may be identified as causes, and the mitigation methods employed by individual pilots. Links to the research on test anxiety will be made.

Recent evidence shows that unresolved interpersonal tensions and intra crew conflicts represent a significant vulnerability for both safety culture and cockpit performance. Multiple operational safety investigations report similar occurrences, confirming that these conflict dynamics are not episodic but structurally embedded in real world flight operations. In a safety critical environment such as aviation, these findings raise important questions regarding the predictive effectiveness of current selection models for cockpit personnel. This exploratory study examines the limitations of widely adopted aviation selection practices and highlights how, despite the integration of technical, cognitive, and relational assessments, they remain only partially effective in identifying behavioural risk factors that may later evolve into operational conflict. Emerging patterns suggest the presence of deeper professional dimensions—often overlooked within traditional assessment frameworks—that contribute to these dynamics and remain insufficiently evaluated in a predictive perspective. The analysis indicates that conflict prone behaviour frequently results from the interaction of several variables: incomplete interpretation of the systemic operational context, suboptimal management of hierarchical relationships, limited behavioural modulation under pressure, and insufficient alignment between individual motivations and organisational goals. Such factors represent one of the least explored and least assessed domains within current selection methodologies. Building on these findings, the paper proposes an evolution of the selection model toward a safety oriented approach grounded in the assessment of systemic compatibility between the candidate and the aviation environment. The proposed model translates ex post evidence from safety investigations into ex ante organisational requirements, targeted assessment procedures, observable behavioural indicators, and explicit thresholds of acceptability aimed at enhancing the accuracy of selection decisions. By extending the traditional purpose of selection—focused primarily on assessing technical professional suitability—with a behavioural risk prediction capability, the model strengthens the identification of individuals capable of operating effectively within high reliability systems. This contributes to consolidating a more robust safety culture and supports a proactive approach to human factors management across aviation operations.

Assessment Centers (ACs) are widely used but resource-intensive procedures for for assessing social competencies personnel selection. Situational Judgment Tests (SJTs) represent an established and more efficient simulation-oriented method for predicting professional success. When developed in a construct-oriented manner, they enable a valid assessment of job-relevant characteristics. In the present study, a construct-based SJT consisting of 12 situations was developed to assess leadership and teamwork for use in the selection of aviation personnel. A scoring key was derived from effectiveness ratings provided by N = 13 experts (aviation psychologists). In a second step, the test was analyzed with regard to its suitability as a selection instrument in the context of a multi-stage selection procedure.
In the first selection stage, N = 172 applicants for the profession of pilot completed various cognitive and knowledge tests as well as a personality questionnaire, and the newly developed construct-based SJT. In the second selection stage, N = 82 applicants, who had passed stage one, participated in an AC consisting of various simulation-based procedures, assessing dimensions of leadership, cooperation, and communication (among others). A subsample completed a work sample test (flight simulator task, N = 41) in the third selection stage and a final interview (N = 33).
The SJT results correlated significantly with different ratings of the AC observers: positive relationships were found for communication, teamwork, and leadership. Furthermore, the SJT-scores showed positive relations with the flight simulator performance and the final interview. Additionally, applicants who passed the cognitive test battery (stage 1) scored significantly higher on the SJT than those who failed. This confirms previous research that found positive relationships between knowledge-based SJTs and cognitive abilities, as SJTs require cognitive processing. Moreover, only two marginal relationships were found between personality scales and the SJT results.
Overall, the observed associations provide first evidence for hypothesis-consistent convergent relationships of the newly developed construct-based SJT with the existing selection procedures. The finding suggests that the SJT validly captures relevant behavioral competencies and represents a promising and efficient pre-selection-tool for aviation personnel selection.

Competency-based assessment frameworks have become a cornerstone of contemporary pilot selection and training systems, particularly within evidence-based training (EBT) models endorsed by international aviation authorities.
This study examines how nine core pilot behavioral competencies—aligned with EASA-recommended competency frameworks—are perceived by different aviation stakeholder groups in terms of both their relative importance in CADET candidate evaluation and their perceived trainability. Specifically, the research compares perceptions among several groups: competency-based interview (CBI)-trained airline pilots, non-CBI-trained airline pilots, aviation psychologists, human factor specialists, and human resources professionals working in aviation organizations.
A cross-sectional relational survey design was employed. The study aims to recruit approximately 120 participants (n≈30 per group) using purposive sampling. Data are collected through a structured questionnaire developed by the researchers, consisting of Likert-type scales assessing: (1) the importance of nine core behavioral competencies, (2) the importance of their respective sub-dimensions, (3) the perceived trainability of each competency, and (4) the perceived trainability of sub-dimensions.
The study seeks to generate a comparative ranking of competencies across groups, identify statistically significant intergroup differences, and explore alignment or divergence in perceptions regarding developmental potential.

The study is currently in the data collection phase. Final empirical findings, including comparative competency rankings and statistically significant intergroup differences, will be presented at the conference.

Personality assessment has always been intuitively relevant and important in fitting a person to a job or selecting for likelihood of success in job performance. However, there are many existing challenges in personality testing such as tools are transparent and easy to fake, measuring intra-individual variance in personality-linked behaviours across different situational contexts and aspects of personality that do not easily lend themselves to self-report survey type tools such as integrity. All these challenges limit the validity and usefulness of many current personality tools in the market. This presentation covers the development of a game-based personality tool designed to 1) reduce faking, 2) tap personality nuances in different work contexts to increase the richness of the assessment and 3) measure aspects of personality that have been more difficult to assess thus far. Combining robust psychological design and data-science capability, this tool aims to generate a unique personality profile tailored to each individual and an assessment of match to specific job roles. The personality dimensions measured by this tool were validated against established measures of the Big Five, demonstrating that the tool captures core personality domains while preserving its situationally grounded and behaviourally expressed framework. The findings support the structural coherence of the underlying trait architecture and indicate that personality expression across varying workplace contexts can be modelled in a systematic and theoretically consistent manner.

Automation in the flight deck, like the Ground Proximity Warning System (GWPS), has substantially mitigated operational risks in commercial aviation and contributed to a reduction in catastrophic accidents. Nevertheless, accident and incident investigations identify persistent human factor (HF) related errors, especially in situations involving mental overload, degraded situational awareness, deviations from standard procedures, and poor crew coordination. These vulnerabilities remain central contributors to safety events such as Controlled Flight Into Terrain (CFIT) and Loss of Control In-Flight (LOC-I).
With recent advances in Artificial Intelligence (AI), cockpit systems may evolve from deterministic automation toward adaptive decision-support acting as cognitive team members that could enhance human performance and team coordination. However, due to the limited assurance of current AI systems and the strict safety requirements under European Union Aviation Safety Agency (EASA) regulation, it becomes necessary to assess the extent to which AI-based systems can mitigate recurrent HF related errors documented in accident and incident reports, without introducing new safety hazards, and to identify operational contexts in which such systems provide measurable safety benefit.
To address this necessity, this paper systematically reviews accident and incident literature to identify recurring HF related safety events and high-risk operational contexts. Based on this evidence, representative use cases are derived in which AI support may mitigate contributing factors to safety events. These use cases are then aligned with the EASA AI guidance material to determine the appropriate level of AI integration that is feasible within regulatory constraints.
Building on this evidence, a conceptual supportive AI system is proposed. The system is designed as a cognitive team partner, incorporating the FORDEC (Facts, Options, Risks and Benefits, Decision, Execution, Check) decision-making framework, that addresses the identified HF risks through structured decision support, dynamic situational modelling, and transparent human-AI interaction, while the human pilots hold final authority.
By aligning accident precursors with operational AI support functions under EASA constraints, a link between the mitigation of safety events and AI system design is established. The resulting conceptual model provides a structured foundation for subsequent empirical validation to assess whether AI-supported teaming can measurably reduce HF related errors in commercial aviation.

In the coming decade, AI-based Digital Assistants (DA) will likely appear in the cockpit and ATC Ops room, carrying out routine tasks or advising pilots during an emergency. The DA concept goes beyond today's Machine Learning-based tools, as the AI-based 'colleague' will engage in dialogue with its human counterparts, leading to the notion of a Human-AI Team.
Aside from technological and regulatory challenges posed by such AI concepts, questions arise of how AI is perceived by people, whether they are ready to embrace it, and how its integration could affect Safety Culture, a core aviation attribute that helps maintain its status as the safest mode of transport.
To address such questions, two surveys were carried out during the HAIKU project, the first to gauge public perceptions of AI in aviation and willingness to accept it (614 participants; 17 European countries), the second to explore the potential impact of AI on aviation safety culture, leveraging realistic simulations with three DA prototypes (21 pilots, 9 ATCOs).
The findings reveal a strong correlation between familiarity with AI and its perception. Individuals with medium-to-high levels of AI knowledge generally see AI as positive. Those with limited or no understanding often perceive it as an antagonist. The AI role as a partner in problem-solving was largely accepted, while AI as final decision-maker was rejected. AI was therefore viewed as a tool rather than a teammate, calling into question the concept of Human-AI Teaming. Furthermore, there was strong consensus on key requirements such as the need for human oversight, management support in case of human-AI performance issues, and dedicated AI training. 'Red lines' concerned unsafe AI actions, the replacement of human staff by AI, and diminished confidentiality of incident reporting.
Overall, the findings at this relatively early stage in aviation AI development indicate favorable conditions for AI introduction, provided humans remain in charge (including being able to switch off the AI). AI literacy for pilots and ATCOs, user-centred design, and a gradual development and validation process involving end users emerge as the critical determinants of successful AI adoption in future aviation systems.

As operational demands are likely to increase in the future, the valid and reliable crew state detection that triggers mitigation functions becomes critical. Mitigations vary from the adaptation of systems in terms of function allocation, task scheduling and content (e.g., abstraction), interaction, to the automated deployment of safety systems.
A crew's functional state mediates performance in relation to psychophysiological costs and depends on human characteristics, current condition, on task demands and on the interaction with the operational environment. Thus, we elaborate on a multimodal approach for assessing and validating a crew's state using a subagent artificial neuronal network. For this purpose, several psychophysiological constructs are described and juxtaposed in a methods box containing four different data sources: Life record data (e.g., recording crew's behaviour in the cockpit), questionnaire data (self-assessment such as perceived workload), test data (objective measurements such as performance) and environmental data (relevant context information such as G-load). The proposed assessment of crew's state detection uses these four sources of data in a multimodal approach and a mathematical verification in a subagent artificial neuronal network (S-ANN). This allows for validation of the crew state detection that is used to trigger adaptive automated functions. Main challenges are coping with limits in predictability and explainability of the S-ANN, considering the inference mechanisms and the model complexity, and elaborating pertinent guarantees on the absence of 'unintended behaviour' of AI-based systems.

In the rapidly evolving landscape of high-reliability industries, the integration of Artificial Intelligence (AI) into aviation represents a pivotal shift in operational safety and human factors history. While technical systems advance at an exponential rate, the successful synergy between human operators and AI assistants will be clarified by understanding of perfectionism and resistance to change that govern acceptance and usage behaviours.

This research explores the "Human-Machine Paradox" by examining the psychological predictors of AI engagement among aviation personnel, focusing specifically on the constructs of Resistance to Change (RTC) and Perfectionism. Resistance to change is conceptualized as a tendency to avoid altering established routines, potentially fostering negative attitudes toward technological disruption. Conversely, perfectionism is treated as a double-edged sword: while high performance standards may motivate the use of AI to achieve high precision, concerns over errors or the perceived loss of operational control may trigger unconscious scepticism toward automated systems.

To investigate these dynamics, the study employs a quantitative, cross-sectional survey design, collecting data from a stratified sample of approximately 250 participants, Cadet pilots and pilots. The survey is composed of resistance to change scales, multidimensional perfectionism scale, general attitude toward artificial intelligence scale. Correlation Analysis will be conducted to identify the strength and direction of the linear relationships between dispositional traits and AI-related outcomes. Difference Analysis (using ANOVA and t-tests) will be employed to uncover significant variations in attitudes and traits between different occupational roles and between active AI users and non-users. Finally, Multiple Regression Analysis will be applied to determine the specific predictive power of RTC and various facets of perfectionism on AI usage attitudes and self-reported usage duration.

The significance of this research lies in its potential to transform how the aviation industry approaches automation. By identifying specific barriers, organizations can move beyond "one-size-fits-all" training and develop personalized Crew Resource Management (CRM) strategies that address the human directly. Ultimately, this study ensures that as we navigate toward an automated future, we preserve the professional identity and psychological safety of the human operators who remain the final authority in the skies.

Reduced-crew and single-pilot operations (RCO/SPO) are investigated as long‑term responses to pilot shortages, economic pressure, and increasing cockpit automation, yet they raise unresolved human‑factors concerns regarding workload, situation awareness, crew coordination, and trust in distributed socio‑technical flight deck architectures. This paper presents a requirement‑driven, certification‑aware development and evaluation of a projection based telepresence system intended to support collaboration between an onboard single pilot and a remote crew operator in RCO/SPO scenarios. Building on CS‑25, AMC 25.1302 and related certification standards, a structured requirements derivation process yielded 50 system requirements, consolidated into 18 key functional and human‑factors requirements that served as evaluation criteria for a weighted Pugh‑matrix comparison of visualization technologies. This process identified a Side Head‑Up Display (Side HUD) concept as the most suitable solution, as it integrates a semi‑transparent NanoAR projection screen into a cockpit without head‑mounted devices, preserving out‑the‑window visibility while enabling visually embodied, bi‑directional collaboration. A prototype architecture combining depth‑camera‑based point‑cloud capture, head tracking, and real‑time rendering was implemented in a fixed‑base cockpit simulator and connected to a ground‑station workplace. In a counterbalanced within‑subjects experiment, 14 participants worked in paired roles across three collaboration modalities: in‑person, audio‑only, and telepresence‑supported interaction during a time‑constrained cooperative task involving instruction transfer, coordination, and execution. Objective measures (instructions executed and communication errors) were complemented by NASA‑TLX (workload), SART (situation awareness), SUS (usability), and semi‑structured interviews. Preliminary results show that telepresence substantially reduces communication errors relative to audio‑only and yields intermediate workload with lower variability, indicating more stable cognitive strategies. SART scores demonstrate improved shared understanding and anticipation compared to audio‑only, particularly for the remote role, while SUS ratings confirm that the Side‑HUD implementation is perceived as clearly more usable than audio‑only collaboration. Although telepresence does not fully replicate co‑located crew interaction, the findings provide empirical evidence that visually embodied, non‑wearable telepresence can mitigate key human‑factors penalties of distributed crew concepts and represents a promising human‑centred enabler for future RCO/SPO flight deck architectures.

General Aviation (GA) cockpits in the United States have seen technological improvements in recent years. However, existing technologies such as a Heads-Up Display (HUD) to provide advanced situational awareness during visual and instrument procedures are essentially non-existent even though the technology has been available to commercial and military platforms for decades. Additionally, general aviation operations typically operate with only one certificated pilot in the cockpit. Many factors contribute to general aviation airplanes and their pilots being more susceptible to fatalities following a loss of situational awareness. These include severe weather and the general operating environment of single pilot flights. Also, the aircraft regulatory environment might prevent permanently installed technologies to be installed unless more thoroughly tested.

This study gave pilots of non-commercial general aviation aircraft access to Extended Reality (XR) glasses connected to an Electronic Flight Bag (EFB) such as Foreflight for measuring pilot perception of improvement in situational awareness. Extended Reality (XR) glasses provide a semi-transparent overlay that allows the pilot to see both the normal cockpit and a virtual Heads-Up Display. The information from Foreflight displayed altitude, indicated airspeed, groundspeed, distance to next waypoint, estimated time enroute, and more. Pilots went flying in simulated instrument conditions and did normal instrument procedure maneuvers. Afterwards, they put on the Extended Reality (XR) glasses providing the HUD and were asked to perform more maneuvers and instrument approaches such as RNAV GPS and ILS approaches.

The researchers debriefed the pilots on their experience with the Extended Reality (XR) glasses. Questions were asked about their perceptions of effectiveness of the technology, ease of use, and the perceived advantages and disadvantages over traditional general aviation cockpit instruments. The researchers reviewed footage from inside the cockpit and utilized ADS-B data to examine the flight paths taken by the pilots using the additional overlay technology. Data analyzed reveal promising paths forward for the blending of Extended Reality (XR) glasses in general aviation cockpits.

Introduction: Stein (1992) investigated scan patterns of licensed air traffic controllers while managing air traffic in a simulated terminal radar environment. The findings revealed that ATCOs require 5-10 minutes to develop situational awareness, after that the scanning patterns stabilised and remained consistent through the control period. Qualitative analysis of scan plots suggested that scanning strategies are influenced by airspace geometry, traffic flow, operational demands and experience/length of service. Clung & Kang (2016) defined and developed new concepts to systematically filter visual scan paths with linguistic inputs from self-reported strategies. They conceptually described scan paths as being circular, linear, trajectory, regional, augmented, proximity-based, density-based or a mix. However, they highlighted that classifying scan paths into shapes is complex and time-consuming and relies on the assumption that an ATCO intends to use a particular strategy. The authors suggest some relevant definitions to filter complex scan paths into manageable forms that can lead to representation of strategies. Definitions include raw scan paths, initial and fundamental global and local scans and aircraft comparisons.
Method: This explorative simulation study investigates whether a visual scan path pattern can be identified by tracking licensed radar controllers across different experience groups, positions and traffic density (sector load). A total of 70 eye tracking datasets are available for analysis. Eye tracking data were recorded using Viewpointsystem (VPS) 19 glasses. Simulation scenarios lasted between 20 to 50 minutes. Data were analysed using VPS fact finder software and standard psychometrics to check for statistical differences.
Expected Results: It is anticipated that distinct visual scanning strategies will emerge as a function of controller experience (novice versus expert), sector traffic load (low, medium, high), and operational role (executive versus planning controller). However, it is not expected that individual scanning paths can be reliably clustered into discrete visual strategies.
Practical conclusions: The results can inform the development of improved training methods, interface design, and automated tools to support controller performance in increasingly complex air traffic environments.

Remote sensing (i.e., contactless) offers significant advantages over contact-based approaches, particularly in safety-critical and high-stakes domains. Unlike wearable sensors which can be cumbersome, distracting, uncomfortable and may interfere with task execution when worn by operators, contactless systems enable unobtrusive, long-term monitoring and facilitate the deployment of multimodal sensing setups contrary to wearable sensors that compete for limited body locations. Recent progress has been driven by advances in various remote sensing technologies, including RGB and infrared cameras, radar systems, and thermal imaging, as well as by data-driven models capable of integrating multimodal physiological and behavioral signals. This narrative review summarizes recent advances (2016–2026) in remote human physiological sensing across safety-critical and high-stakes application domains: automotive, aviation, industrial, military, and healthcare settings.

Safety-critical domains seek reliable contactless monitoring techniques of fatigue (Tan et al., 2025), stress (Gioia et al., 2022), situation awareness (Li et al., 2023), startle responses (Schwartz et al., 2025), or cognitive workload (Causse et al., 2024; Vora et al., 2017). To support these applications, multimodal platforms integrate diverse sensors for example RGB and infrared cameras to capture blinks, skin-color or thermal shifts for respiration (F. Yang et al., 2022), heart rate or stress detection etc.; radar (including millimeter-wave systems) to capture mechanical chest displacement for the respiration (Z. Yang et al., 2025); remote eye tracking can monitor pupillary dilation and gaze patterns as proxies for cognitive workload (Gorin et al., 2024); and sometimes joint audio for analyzing the vocal strain (Van Puyvelde et al., 2018). When remote sensors are combined with deep-learning models operating on synchronized multimodal streams, physiological signal extraction from heterogeneous and noisy data can be possible (Kong et al., 2024; Shen et al., 2025). Persistent challenges include motion artifacts, environmental variability, inter-individual calibration, privacy and ethical concerns. Future work emphasizes adaptive human–machine interfaces, explainable AI, and standards for safe and responsible deployment.

Work sample tests (WST) play a vital role in ab initio air traffic controller (ATCO) selection. They measure complex performance and are developed based on demands of the working position. While ability tests focus on the measurement of isolated abilities, WST require the application of a specific range of abilities. They are usually closely designed to meet the core aspects of the job and thus offer high face validity. Traditionally, the multidimensionality of the measurement requires human judgements and ratings to adequately assess the individual's performance. At the same time, the deployment of human ratings within a selection process is time and cost intensive. Furthermore, raters need intensive initial training and ongoing standardization training to reach high levels of interrater reliability. Automatic scoring presents an opportunity to assess the performance in WST with high effectivity and efficiency.
We present a practical use case from the DLR selection of ab-initio air traffic controllers regarding an ATC-specific WST. The WST simulates an ATCO radar working position and requires multitasking and the multimodal integration of information. An automated scoring system for this WST was conceptualized and developed: The preprocessed logfile data of a sample of N = 4046 applicants from 7 years of selection and the corresponding results from expert ratings of performance were used as training data. Regression models were trained and optimized. Subsequently, the resulting scoring algorithm was tested with an independent validation sample of N = 711. The scoring algorithm performed very well in predicting the human expert ratings and will be implemented in the selection process soon. We will present empirical results of the validation sample, showcase the implementation process and will address the challenges and opportunities of our approach.

Public discussion (e.g. Microsoft attention spans report, 2015) and ATCO training instructor's gossip regularly stress issues around a loss of "attention span" compared to older "generations" and claim a problematic development of mental performance of young people. Often these are attributed to the use of modern communication devices and media. However, there is few empirical evidence to demonstrate this, apart from e.g. reports on what is called the Flynn or reverse Flynn effect, which however do not report consistent and unique effects (cf. e.g. Oberleiter et al., 2024, Nordmo et al., 2024).
The paper reports findings based on comparable psychometric test data from 25.000 applicants of an online selection process for an ATCO training participating between 2017 and 2026. Data on development over time, construct validity (e.g. intercorrelations between mental performance domains such as GMA and attention), relations with moderating factors such as school performance, gaming and online activities are reported and, where appropriate, compared with expected effects e.g. reported in the test manuals or other literature.
Based on further data from a classical concentration test, the methodological challenge of comparing concentration test performance across generations is discussed and showcased based on exemplary samples (respective N = 50 to 120) across 25 years.
Results show more continuity than change in mental ability level of ATCO training applicants. Regarding concentration span and performance, it can be concluded that a reliable and valid comparison of actual concentration test performance across the last generations can hardly be achieved due to significant changes in test formats and media, sampling effects etc. Implications for aviation selection and training are discussed.

The training of ATCo's (Air Traffic Controllers) requires the development not only of technical competencies but also of non-technical skills such as situational awareness, communication, stress management, and teamwork. In recent years, experiential learning methodologies have increasingly been adopted in aviation training programs, as they have proven to be particularly suitable for increasing emotional awareness and self-efficacy in participants.

This paper examines the benefits of the experiential approach in the development of non-technical skills among air traffic controllers, linking these benefits to feedback collected from participants during training events lasting between 6 and 8 hours, integrated with pre-training and post-training questionnaires.

The collected data come from different kinds of training events on ATCo's courses reflecting a wide range of ATCo attendant feedbacks (civilian, military, ab-initio ATCo's, expert ATCo's, training for specialization or attending a refreshing course).

Qualitative results indicate a significant increase in both operational and personal awareness, particularly when the debriefing phase was explicitly linked to the operational context of air traffic control and when the initial activities facilitated the emergence of the main psychological and emotional dynamics experienced by participants.

In recent applications designed to increase system resilience through teambuilding and communication awareness of this training model, the target audience has been expanded beyond ATCo's employed at neighboring and hierarchically related unit (e.g., APP and TWR) to include pilots representing airlines serving those ATC environments. The results, although requiring further development, further support the view that this type of training tool is effective in enhancing system resilience and strengthening safety.

Extended Reality (XR) technology, encompassing both Virtual Reality (VR) and Augmented Reality (AR), has seen rapid development in recent years. This offers new training possibilities for aviation professionals. In air traffic control (ATC), training is traditionally resource-intensive and requires access to real or highly realistic operational environments, particularly regarding control tower operations. XR presents an opportunity to provide immersive, flexible, and cost-effective training scenarios.

A study with 24 test persons evaluated the usability of two state-of-the-art XR Head-Mounted Displays (HMD) in a control tower environment. The study did focus on comfort, wearability, interaction with ATC equipment, peripheral vision, spatial awareness, and responsiveness. The 24 participants consisted of control tower Air Traffic Controllers (ATCO) and ATCO trainees. They completed static and simulation tests across three setups: A baseline setup with conventional screens and no HMD, a setup A which used the Apple Vision Pro M2 and a setup B with the Varjo XR-4. Setup A featured a special head-mount, which omitted the standard Vision Pro seal around the eyes. This allowed a direct head-down view of the flight strip board and the desk. In setup B, the Varjo XR-4 sealed the entire view. The flight strip board and the desk were only visible through the cameras of the HMD.

A mixed-methods approach combined quantitative tests (e.g., simulator sickness and reaction time tests, comfort scales, readability assessments) with a qualitative user experience evaluation.

Compared to the baseline setup, setup A showed unchanged reaction times but slightly higher mental demand and mild sickness within acceptable limits. Setup B produced longer reaction times, greater sickness, reduced comfort and a more difficult hand–eye coordination, all together resulting in a lower acceptance rating. Thus, while setup A deemed to be usable for ATCO training, setup B negatively affected all evaluated dimensions.

Overall, the results indicate, that an XR HMD could be a valuable supplementary training tool for Air Traffic Controllers (ATCO). However, the device has to be carefully selected and training objectives have to be carefully planned an validated.

The integration of artificial intelligence (AI) into the flight deck requires a fundamental rethink of how authority is distributed between pilots and automated systems. While current regulatory discourse—most notably from EASA (2023) and the EU AI Act (2024)—stresses the need for "human-centric" systems, these frameworks remain largely theoretical. A critical gap remains in defining and operationalising human cognitive authority within high-stakes environments. This paper addresses this by proposing a measurable framework for cognitive authority, centred on a pilot's functional capacity to understand, challenge, and override AI-generated outputs while navigating time pressure and systemic uncertainty.

By analysing Endsley's (1994) situation awareness model with Dekker's (2003) work on resilience and ICAO's competency-based training standards, I conceptualise cognitive authority through three interconnected lenses: (1) attentional alignment—the synchronicity between pilot gaze and AI saliency; (2) interpretive transparency—the pilot's grasp of underlying system logic; and (3) override readiness—the behavioural capacity to counteract AI guidance.

The proposed framework moves beyond theoretical discourse by integrating physiological and behavioural data, including eye-tracking latencies and challenge–response patterns, within an Evidence-Based Training (EBT) framework. Drawing on neuroergonomic findings (Causse et al., 2011; Peysakhovich et al., 2018), the study utilises simulator-based "authority stress scenarios" to pinpoint where measurable degradation in human oversight emerges.

The contribution of this work is twofold. First, it provides regulators with an empirical toolkit to move beyond normative "trustworthy AI" rhetoric toward verifiable safety standards. Second, it offers operators a pathway to evolve recurrent training by implementing cognitive authority metrics directly into flight crew assessment cycles to mitigate out-of-the-loop effects associated with advanced flight deck automation.

The introduction of artificial intelligence in complex work environments, such as air traffic management, raises new transparency, acceptability and teaming difficulties in human-system interaction (Xia et al., 2025). While digital assistants are increasingly attributed agentic capabilities (e.g., the ability to set goals, plan action, or make decisions to achieve those goals), considering them as a team member in joint activity remains a challenge (Zhao et al., 2025). Mutual predictability has been proposed as a key prerequisite for effective human-agent teaming (Klien et al., 2004). To date, however, research has largely focused on improving the human operator's understanding of artificial agents (AA) through transparency and explainability, while the AA's ability to anticipate human operators' intentions has received substantially less attention. To address this limitation, this study employs a multi-stage human-centric methodology to model Air Traffic Controllers' (ATCOs) goals when responding to pilot requests. First, a questionnaire (N = 55 ATCOs) identified three critical use cases where AA would most benefit safety, operational efficiency, and ATCO workload: direct routing, level change, and weather avoidance requests. Second, more than 15 hours of semi-structured interviews (N = 6) were conducted to perform a Hierarchical Task Analysis (HTA) for each use case, which aims to identify what an ATCO is required to do in terms of actions and/or cognitive processes to achieve a system goal (Kirwan and Ainsworth, 1992). HTAs allow activity to be broken down into goals, sub-goals and actions, as well as the plan (order) in which they are performed (Stanton, 2006). Finally, the resulting HTA models were cross-validated through additional ATCO interviews (N = 3) and an online validation survey (N = 21). By mapping out how controllers prioritize their tasks, these task models allow us to design AA that actually understand the operator's plan. Ultimately, this project aims to use these HTAs to inform a digital assistant with ATCO goals, supporting the anticipation of operator needs and providing relevant information when responding to the selected use cases. Identifying operators' intents and goals is crucial for Hybrid Human-AI decision making. We will further test this hypothesis in human-in-the-loop simulations with en-route ATCOs.

The EU Digital European Sky programme assumes a digital transformation where human-artificial intelligence teaming (HAIT) is a foundation.

The theoretical framework that connects humans and technology is automation. Is automation an appropriate framework to design work systems within the scope of this digital future? Where such designs need to navigate interdependence between human actors and machine agents in HAIT?

These questions are the subject of this abstract drawing upon experience from participation within research projects undertaken through the EU's SESAR programme This experience within these projects reveals that the design of 'work systems' tends towards amplifications of extant concepts and methods of operation as opposed to exploiting the potential that HAIT holds.

Envisaged work systems build on models that are commonly derived through a task model of ATCOs. This forms the basis of function allocation consistent with extant automation philosophy. Routine tasks for example, but also complex non-routine tasks, such as conflict detection and resolution.

A consequence of this is that HAIT is influenced and limited by the automation framework. HAIT is different from automation we contend. Designs exploiting the potential of joint activity through automation and function allocation fail to transcend 'MABA-MABA' lists.

As an example, the most common operating team structure in en-route ATC in Europe. is a two-person team consisting of an Executive and Planner controller. Each sector team functions in ways that are mutually supportive, dependant,, but also independent that provides a subtle and functionally adaptive choreography in sector operations. Are machine agents introduced into this team structure as support to each role? Where this has been adopted, it is evident that new and changes coordination costs have been introduced through limitations of the new HAIT structure. Radical changes of roles out with the automation framework led to different structures.

Seen through a lens of cognitive systems engineering, HAIT is a technology where, exploiting the technological potential, articulated through the use of the automation metaphor, skews and limits the design. As opposed to exploiting the potential of joint activity and fundamentally different paths for coordination and interdependence because of the opportunities of HAIT transcend.

Effective teamwork is essential for safety and operational success in aviation. Decades of psychological research has focused on understanding and supporting human–human teamwork, resulting in well-established models such as Salas et al.'s (2005) Big Five and Crew Resource Management (CRM). The future integration of AI into operational environments (cockpits; air traffic control centres), nominally called Human-AI Teaming (HAT), exposes a growing tension between established, human-centric models of teamwork and the reality of human-AI interaction. This paper explores the basis of HAT in the context of two case studies (civil and military aviation), aiming to determine if HAT is a good idea, and if so, what 'good' and 'bad' HAT would look like.

In earlier work on human–AI interaction this development was framed as a transition "from tool to teammate." This premise has since been widely challenged, with critics arguing that treating AI as a teammate risks anthropomorphism, inflated expectations, and unsafe reliance. Yet, simply rejecting the notion of teaming does not resolve the problem: humans are already interacting and collaborating with AI systems in ways that resemble coordination with other agents, whether or not the term "teammate" is theoretically justified. In safety-critical domains such as aviation, this creates an urgent need to understand how fundamental differences between humans and AI shape interaction.

Four such fundamental differences, along with their implications for human–AI collaboration, are identified: (1) emotional and social intelligence, influencing decision-making and social interaction; (2) cognitive resources, including model alignment, cognitive load distribution, and temporality; (3) embodiment and physicality, shaping perceptions of risk, vulnerability, and trust; and (4) capacity for moral and ethical reasoning, with consequences for accountability, ethical decision-making, and oversight. For each dimension, potential benefits ("the good"), risks ("the bad"), and open questions that remain insufficiently understood ("the unknown") are identified and contextualised via the two use cases. The results point to the most promising avenues, potential 'dead ends' to avoid, and critical research needs to deliver safe and effective human–AI collaboration in aviation.

The psychological demands placed on pilots are increasing as aviation systems become more complex and operational environments more dynamic. Developing adaptive capacities early in a pilot's career may therefore play a critical role in sustaining both performance and wellbeing. Psychological resilience: commonly defined as the capacity to adapt positively and recover from stress, has been widely studied in high-reliability professions, yet empirical research examining resilience within ab-initio pilot populations remains limited. This study investigated psychological resilience among student pilots undergoing professional flight training and explored its relationship with perceived stress and mental wellbeing.
A cross-sectional survey was conducted with 123 ab-initio pilots enrolled in two integrated airline pilot training programmes (ATPL and MPL). Participants completed three validated psychological instruments: Brief Resilience Scale (BRS), Psychological General Well-Being Index – Short Version (PGWB-S), and Perceived Stress Scale (PSS-10). Additionally, qualitative open-ended questions, gathered contextualised perspectives on the constructs lived experiences by the students. Descriptive statistics were compared with available normative data, while inferential analyses were conducted to explore relationships between psychological variables and different training characteristics.
Quantitative findings revealed that ab-initio pilots demonstrated moderate levels of psychological resilience and wellbeing, with perceived stress increasing across training phases. Resilience was positively associated with wellbeing and negatively correlated with stress, indicating its moderating role. Differences in resilience and stress patterns were also observed across training phases, highlighting the dynamic psychological demands encountered during different stages of pilot development. Prior flying experience and training phase influenced psychological functioning, while programme type did not. Qualitative insights reinforced these patterns, highlighting the importance of resilience, social support, and coping strategies, while also revealing systemic stressors and yet to be met support needs.
These findings emphasise the importance of psychological resilience as a protective factor supporting adaptation in demanding aviation training environments. Given the results, it is suggested that resilience-focused educational and developmental interventions should be invested to assess the impact on trainee pilots' capacity to manage stress and maintain wellbeing during training. From broader aviation psychology perspective, integrating resilience development within training and support frameworks may contribute to strengthening human performance and long-term workforce sustainability in aviation.

Background:
As EBT and CBTA approaches transform the behavioural and attitudinal demands placed on aviation instructors, systematic assessment of instructor behaviour has never been more relevant. Yet it remains largely absent from the empirical literature, with selection and assessment activity historically prioritising pilot/ATCO trainees. Addressing this gap, this study describes the development and exploratory validation of a behavioural assessment tool capturing individual differences in instructional approaches, exploring relationships between these approaches and perceived teaching effectiveness, and examining aviation-specific behavioural patterns via a non-aviation comparison sample.

Method:
The assessment integrated two established inventories, Grasha's TSI and Mosston and Ashworth's Spectrum of Teaching Styles, selected for their relevance across theoretical and practical skills-based instructional contexts. Teaching effectiveness was assessed using competency-aligned frameworks, drawn from IATA Instructor Competencies (aviation) and Teaching Fundamentals (non-aviation). Following item refinement supported by SME review and a pilot trial, 61 participants completed the assessment. Analyses examined scale reliability, group differences, and relationships between behavioural dimensions and effectiveness ratings.

Results:
Internal consistency of scales was acceptable after item refinement (Cronbach's α = .722–.831). Significant differences in dominant behavioural approaches emerged between aviation instructors and their non-aviation counterparts, with several dimensions demonstrating significant correlations with teaching effectiveness ratings. Exploratory regression modelling found instructional approaches accounted for 56% of variance in performance ratings, explaining an additional 28% of variance above and beyond demographic factors and occupational group. A small number of instructional approaches emerged as significant, independent predictors of performance ratings – both positive and negative – suggesting both the presence and absence of particular behavioural tendencies influence teaching effectiveness.

Practical Implications:
A follow-on developmental workshop demonstrated that individual feedback reports supported structured reflection on instructional tendencies, peer-led discussion, and identification of targeted behavioural adjustments. These findings highlight the assessment's utility as a facilitated developmental tool for instructor training, CPD, and implementation of EBT.

Limitations and Future Frontiers:
Limitations include the modest sample size, cross-sectional design, and reliance on self-rated measures. Future research should prioritise larger samples incorporating objective performance criteria, and longitudinal validation. These early findings nonetheless suggest meaningful potential for behavioural assessment within instructor development and continuing professional practice.

Recurrent pilot training is central to maintaining operational competence and safety across the pilot's professional lifespan. While Competency-Based Training and Assessment (CBTA) frameworks have strengthened performance evaluation, recurrent training structures often remain temporally compressed, cognitively dense, and insufficiently differentiated across career stages. As a result, engagement, motivation, and long-term progression across the pilot career lifecycle are being treated as secondary outcomes rather than intentional design variables within the training system.
This paper presents a Human Centric Model developed through exploratory MSc research in Aviation Human Factors, awarded Distinction. The research examined how principles from elite sport, particularly structured developmental cycles, calibrated cognitive load, consolidation phases, and sustained motivational strategies, can inform the evolution of recurrent pilot training within existing CBTA frameworks. Drawing on cognitive load theory, organizational and instructional climate research, engagement literature, and qualitative data from expert interviews and practitioner questionnaires, the model conceptualizes recurrent training as a layered and system-centric developmental process rather than a discrete evaluative event.
The model integrates three interdependent layers. First, organizational and instructional climate, including leadership behaviours, psychological safety, and instructor as coach orientation, establishes the conditions under which adaptive learning can occur. Second, structured cognitive challenge is calibrated to experience level and professional phase, ensuring appropriate difficulty without cumulative overload. Third, embedded opportunities for reflection and consolidation support knowledge integration, self-regulation, and progression across the pilot career lifecycle.
Sustainable performance is conceptualized as a system-level outcome. Pilots, instructors, and organizations must each feel empowered and psychologically safe to engage in development-oriented training processes. The paper further explores how this human-centric model may be layered into existing Safety Management System architectures, aligning recurrent training more explicitly with organizational safety strategy. In the context of increasing automation and AI-supported operations, the model also provides a framework for calibrating cognitive load and trust development within human AI teaming environments, ensuring that technological advancement is accompanied by deliberate human performance design.

Continuous change is inherent in airline training and operational processes. When procedural or technological updates directly affect daily operations and training practices, their effective implementation becomes critical. Ineffective transition processes may cause operational disruptions, reduce training efficiency, increase training costs, and introduce risks that may compromise flight safety. Beyond economic implications, such disruptions may weaken safety margins in high-reliability environments. Accordingly, this study examines the barriers that emerge when previously learned knowledge and routines must be abandoned and explores strategies to support safer and more effective adaptation. Within the literature, unlearning is conceptualized as an intentional process through which individuals or organizations consciously discard established knowledge structures, routines, and beliefs. Drawing on organizational learning and cognitive psychology perspectives, unlearning involves questioning outdated practices, reassessing existing cognitive schemas, and developing alternative behavioral patterns that facilitate adaptation to new conditions.
This qualitative study draws on semi-structured interviews conducted with 21 commercial airline pilots employed by an aviation company in Turkiye, with an average flight experience of approximately 6,500 hours. Data were analyzed using MAXQDA through reflexive thematic analysis supported by systematic coding procedures.
Five interrelated themes emerged. (1) Habit Persistence reflects the tendency of deeply ingrained motor and verbal routines to resurface automatically after procedural changes. (2) System-Induced Confusion captures difficulties arising from aircraft variants, interface similarities, and technological changes that increase cognitive load. (3) Training Transfer Challenges highlight limited practice opportunities, fidelity gaps between simulator and real operations, and weak consolidation of newly learned procedures. (4) Stress-Driven Reversion demonstrates how time pressure, perceived risk, and the need for psychological comfort trigger fallback to familiar routines. (5) Organizational Adaptation Barriers reveal inconsistencies in SOPs, instructor standardization, communication of change rationale, and operational performance pressures that complicate procedural change.
Findings indicate that unlearning barriers emerge from the interaction of cognitive automaticity, system design complexity, emotional responses under operational stress, and organizational implementation practices. Addressing these barriers requires training designs that strengthen transfer and adaptability, human-centred procedural communication, and organizational strategies that support proactive unlearning. Promoting deeper unlearning processes may enhance strategic flexibility, reduce error potential, and strengthen the safety resilience of aviation systems.

Military training emphasizes orderliness, discipline, stress regulation, and mental resilience. Many recruits undergo such training during young adulthood, a period associated with ongoing personality maturation. These experiences may be associated with differences in how individuals perceive and report personality characteristics.
Within the Royal Netherlands Air and Space Force (RNLASF), big five personality traits are assessed as part of the pilot selection process, with particular focus on conscientiousness and neuroticism. Since 2023, people currently serving in the military have been allowed to apply for pilot training alongside civilian applicants, enabling direct comparison within a single selection system. The present study examined whether candidates currently serving in the military differ in personality profiles from civilian candidates and to what extent age accounts for such differences. NEO-PI-R data from 1,523 pilot candidates were analyzed.
For conscientiousness, military candidates scored higher than civilian candidates, t(303.40) = −3.02, p = .003, d = 0.2. For neuroticism, the effect was in the opposite direction, with military candidates scoring lower than civilian candidates, t(1520) = 4.20, p < .001, d = 0.34. No significant group differences were found for extraversion , openness or altruism. Age was positively associated with conscientiousness (r = .17, p < .001) and negatively associated with neuroticism (r = −.17, p < .001). After controlling for age, the group difference in conscientiousness was no longer significant, indicating that this effect was largely age-related. In contrast, for neuroticism, a main effect of candidate group persisted after correcting for age (B= −22.53, p= .009), alongside a significant main effect of age (B= −0.89, p < .001). A significant age × candidate group interaction (B =0.81, p =.023) indicated that neuroticism decreased with age among civilian candidates, whereas this decrease was substantially attenuated among military candidates.
Overall, the findings demonstrate small but systematic personality differences between military and civilian pilot candidates that are not fully explained by age. These results suggest that the interpretation of personality scores in military pilot selection may require consideration of prior military service and age when applying a single normative framework in military aviation selection contexts.

Background. A game-based mobile application was developed and trialed for pilot recruitment and selection. As opposed to traditional selection tests, the game-based assessment was designed to be deeply immersive and entertaining first-person combat flight simulator game, available as a mobile application to assess cognitive performance under dynamic and time-pressured conditions. This study aimed to investigate the psychometric value of the game-based assessment.

Research Questions. To evaluate the validity of the tool, the study examined (1) its construct validity in measuring critical attributes required of a pilot (e.g., psychomotor ability, multitasking performance, and information processing) and (2) its predictive ability to pilot performance.

Methods. Pilot trainees who had successfully passed the Republic of Singapore Air Force (RSAF)'s initial selection phase were invited to participate in the game-based assessment. Convergent validity will be examined by comparing game scores against established computerised assessments administered at the initial selection phase. Regression analyses will be conducted to evaluate the extent to which the game-based assessment predict the subsequent screening and training outcomes.

Results. Preliminary analysis showed that the game-based assessment correlated strongly with factors derived from confirmatory factor analysis of the established computerised assessments. All challenges in the assessment were significantly correlated with psychomotor and multitasking ability in the established tests. Challenge 3 and 5 which included calculation tasks were significantly correlated with established tests measuring symbolic reasoning.

Implications. These findings provide preliminary support for the construct validity of the game-based assessment measuring critical attributes required for a pilot. Its alignment with established computerised assessments suggests potential utility as an early-stage assessment tool within pilot selection systems. As an accessible and immersive mobile platform, the tool may optimise selection efficiency while also enhancing candidate engagement and potentially widening the talent pool without compromising its validity.

Aviation environments necessitate the continuous processing of dynamic information under time constraints, demanding persistent situational awareness and rapid and accurate decision-making. Consequently, cognitive skills including perception, attention, working memory, spatial reasoning, and multitasking have been recognized as essential competencies for pilot performance. Previous studies demonstrated that cognitive test performance is among the strongest predictors of pilot training success and operational performance (Goeters et al., 2004; Martinussen, 1996; Maschke et al., 2011; Matton et al., 2020). Despite this evidence, many traditional selection tools remain either insufficiently aligned with aviation-specific task demands or lack contemporary psychometric validation. To fill this gap, the present study introduces the Psychometrics and Assessment Center Exercises (PACE), a multidimensional cognitive battery specifically designed for pilot selection. There are nine subtests in PACE that test working memory (audio-visual and verbal memory tasks), attention (sustained attention and vigilance tasks), spatial reasoning (3D spatial perception and spatial orientation tasks), math and physics skills, and psychomotor coordination with multitasking. Each subtest used a scoring algorithm that considered accuracy, response time, and error rates for each task.
The study comprised 2505 ab initio pilot candidates (ages 22–30) selected via a national recruitment program. All participants completed the PACE; assessments were administered under standardized conditions. Convergent and discriminant validity were examined, and confirmatory factor analyses (CFA) tested the latent structure. Internal consistency was evaluated using Cronbach's alpha and split-half reliability. Furthermore, normative data were established for each individual subtest, ensuring standardized interpretation of candidate performance across domains.
Results demonstrated satisfactory convergent validity, with significant correlations observed between tests measuring the same constructs within the domains of working memory, spatial reasoning, attention and multitasking. Cross-construct correlations were weak, supporting discriminant validity. CFA supported a higher-order general cognitive ability factor. Reliability coefficients further resulted in medium to strong internal consistency in cognitive abilities.
Overall, our findings indicate that PACE is a psychometrically valid aviation-specific assessment battery capturing both general cognitive ability and domain-relevant competencies. Our results indicate that its incorporation into pilot selection systems may enhance predictive accuracy, improve training efficiency, and ultimately contribute to long-term aviation safety.

Offshore helicopter operations involve demanding flight conditions and high responsibility for crew and passengers' safety. In response to several aviation safety events, the European Union Aviation Safety Agency (EASA) introduced regulatory guidance emphasising the importance of psychological aspects in aviation safety. It highlights the relevance of psychological assessment in identifying factors that may affect safe flight operations. Consequently, evidence-based approaches and modern methods of pilot selection to understanding the psychological and operational requirements of offshore helicopter pilots are essential for supporting effective pilot selection maintaining a strong safety culture.

A job analysis was conducted to develop a requirement profile for offshore helicopter pilots combining a review of operational documentation and regulatory guidelines with structured expert interviews involving experienced offshore pilots, instructors, as well as training, safety, and flight operation managers. The results were summarised as 'POWER' requirements, which served as a basic pilot profile.

These competencies include strong English language proficiency, high motivation, willingness to learn and comply with standard operating procedures, as well as flexibility and an open mindset in dynamic operational contexts. In addition, proactive problem-solving, positive thinking, and high adaptability were identified as essential characteristics for managing complex offshore conditions. Social competencies also emerged as critical, particularly effective communication and teamwork, high social awareness, patience, and respectful interaction with colleagues from diverse cultural and religious backgrounds.
Based on the derived requirement profile, a set of psychological assessment instruments was selected and implemented to evaluate the identified POWER profile in ready-entry pilots. The assessment covered relevant domains such as cognitive aspects, personality characteristics, interpersonal competencies, and safety-related attitudes. Data were collected from a sample of ready-entry helicopter pilots (N = 46) regarding different aspects using cognitive tasks, a workload test and personality questionnaires.

Initial analyses indicate that certain personality traits, such as a sense of responsibility, and high cooperative behaviour, as well as a strong commitment to safety culture, are particularly relevant to safe offshore operations. These findings highlight the relevance of the identified requirements and underscore the potential contribution of structured psychological assessments to evidence-based selection decisions intended to promote safety-oriented behaviour in offshore helicopter operations.

Aviation authorities such as the European Union Aviation Safety Agency, have progressively expanded requirements for Human Factors (HF) analysis in the certification of large airplanes. As HF task volume and complexity grow, the aviation industry faces a global HF engineer shortage driven by increasing workload, workforce aging and limited university education in flight deck HF design and certification. This creates demand for new tools that can compensate for this shortage. AI (Artificial Inteligence) could support selected HF activities. This work focuses on deveoplment and evaluation of AI-tool for identification of applicable regulations, a labor‑intensive, cognitively demanding task due to frequent amendments and complex cross‑references across Certification Specifications, Acceptable Means of Compliance and Guidance Material, as well as material from other Authorities (FAA, TCCA, or ENAC) and industry organizations (e.g., EUROCAE, SAE, and RTCA). To address this, the Institute of Aircraft Systems at the University of Stuttgart and Honeywell Aerospace Technologies have developed and evaluated an AI‑based system that produces set of applicable regulatory paragraphs with supporting reasoning. The objective is to reduce HF practitioners' workload and improve consistency of interpretation. Although Large Language Models (LLMs) can reason over such texts, initial evaluation revealed hallucinations and incompleteness in purely LLM‑based solutions. Therefore, we created a hybrid neuro‑symbolic approach combining LLM semantic analysis with an ontology‑based regulatory model encoding structured relationships between regulations, applicability criteria, and system characteristics. The ontology constrains and validates model outputs, improving completeness, traceability, and reliability, positioning the system as decision support rather than a replacement for expert judgment. To evaluate integration into HF engineering workflow, we conducted a within-subject study using three fictional avionics systems as input. For each system, a human HF expert and AI-system independently generated set of applicable regulations. A blinded independent HF expert rated output correctness following a washout period between ratings. The measures included output generation time, expert-rated output correctness and expert correction time. Results provide insight to human performance, AI performance, and human-AI collaboration effectiveness. The paper discusses HF implications, including overreliance, automation bias, deskilling risk, and junior practitioners' development, supporting responsible AI integration into safety‑critical aviation certification environment.

This research explores the challenges of resource dilemmas in airport collaborative decision making (ACDM) on team dynamics. In a laboratory task, triadic teams consisting of a human dyad and an autonomous agent (HAT) are compared with triadic teams consisting only of humans (HUM). We hypothesize that fairness is less emphasized in HATs, potentially leading to AI exploitation.
Each experimental session includes three sequential games, each embedded in its own supervisory-control scenario (SCS). The game models the resource dilemma, where players balance individual gains against team penalties for selfish choices. In each SCS, two players can defect without team penalty. The third agent is simulated by a computer script. The two participants can communicate after each game, aiding self-governance during transitions.
A 2x2 design is employed, varying the third teammate (human or AI) and their decision foci (towards promoting individual gains or preventing collective punishment). We expected that when this agent is biased to refrain from promoting self-interests, this tends to lead to more defections against an AI compared to a human.
Results show that for conditions with a third agent being biased to promote own self-interests, the human decisions actually differ whether an autonomous agent or a third person is supposed. In the first case, the game outcome tend to be a fair distribution of gains among the two participants. In contrast, a conspiracy against the other participant is more frequently observed when a third person is supposed behind the third agent. This, we assume, should induce social friction into the dyad.
However, when the simulated third teammate (regardless whether it is suspected to be a person or an autonomous agent) acts like a "chump" who always chooses the option to prevent group punishment, the participants tend to take advantage of them and usually conspire to maximize their self-interests.
Based on previous evidence, we expect for experimental conditions with higher social friction, there will be lower synchronization of attentional foci during supervisory control. The analysis of attentional synchronizations is ongoing, and the results, as well as their practical implications will be presented at the conference.

The introduction of AI-based cockpit assistant systems offers the potential to enhance workload optimization and operational efficiency, particularly in the context of Reduced Crew Operations (RCO). Pilots are increasingly required to manage complex situations with limited resources, leading to higher cognitive demands and potential bottlenecks in information acquisition and processing. In this context, decision-making becomes a critical factor for safe and efficient flight operations. AI-based systems may support pilots by structuring relevant information, assessing available options and prioritizing them to support decision-making, particularly in complex and time-critical scenarios.
This research paper investigates the pilot's decision-making process - which plays a central role in human-machine interaction - in complex scenarios with and without the support of a Level-1B AI-based system (EASA Roadmap 2.0, 2023). Fundamental research on this topic, based on semi-standardized expert interviews, was presented at EAAP 2024 and provided initial insights into pilots' expectations regarding AI-based decision support.
The present study extends this work by evaluating the AI-based cockpit assistant system in an experimental simulator study. Seven commercial pilots conducted a diversion scenario in a cockpit simulator under single-pilot operations using the FOR-DEC decision-making model. During the scenarios, the test persons were instructed to use the think-aloud method and a secondary task was applied, followed by questionnaires evaluating the usability, workload and trust of the system.
The behavioural data show improvements both in operational efficiency of the decision-making process itself and in the performance of the secondary task. The subjective assessment of the workload yielded inconsistent results, which are discussed in detail. Subjective ratings of usability, trust and added value of the system were moderately high. A qualitative analysis of open comments reveals cautious support of alternative concepts like AI and RCO as a challenge, which will be further discussed in the context of explainable AI and last authority.
The findings contribute to the human-AI teaming research by providing empirical insights into efficiency, workload and trust when using an AI-based cockpit assistant system, highlighting the importance of developing such systems in close alignment with their intended operational context.

Inspection tasks in safety-critical domains such as aviation demand high levels of attention, accuracy, and situational awareness from personnel operating under strict regulatory frameworks. Recent advances in artificial intelligence (AI) and augmented reality (AR) offer new opportunities to support human inspectors by providing real-time guidance, automated detection, and contextual information directly within the inspector's field of view. However, the integration of such technologies into regulated, safety-critical environments, raises important human factors questions regarding trust in automation, cognitive workload, and the conditions under which effective human–AI collaboration can be achieved.
This paper presents findings from the design, deployment and structured evaluation of an AI- and AR-assisted inspection system developed to support human inspectors in aviation operational environments. The system integrates computer vision–based detection with AR visualization to highlight relevant inspection areas and present decision support during inspection tasks. Consistent with a human-centered AI approach, the system is designed to augment rather than replace human expertise, ensuring the inspector remains the primary decision-maker while maintaining human oversight throughout.
Evidence from field evaluations across operational scenarios reveals several key findings. First, AI-generated visual cues can significantly support detection and guide attention but require clear explanations to maintain appropriate trust. Second, AR overlays improve situational awareness when conveying concise, context-relevant information, but excessive or poorly prioritized visual elements increase cognitive workload. Third, inspectors demonstrate greater confidence and sustained engagement when AI outputs are presented as decision support recommendations rather than as automated decisions. Finally, transparency in AI reasoning and the ability for inspectors to confirm or override system suggestions emerge as defining features of effective human–AI teaming. These findings contribute to the understanding of human–AI teaming in aviation inspection contexts and provide evidence-based design recommendations for human-centered AI and AR systems intended for safety-critical operational environments.

This study examined age-related alterations in resting-state brain function in a large cohort of middle-aged to older professional pilots, representing a critical transitional stage of cognitive aging in a high-expertise context. A total of 139 actively flying pilots aged 41–60 years were included, a period characterized by adaptive neural plasticity and early functional reorganization.
To assess age-related changes across multiple levels, we combined voxel-wise measures of local activity with network-level functional connectivity analyses. Voxel-wise metrics included amplitude of low-frequency fluctuations (ALFF), fractional ALFF (fALFF), regional homogeneity (ReHo), and degree centrality (DC), capturing complementary aspects of spontaneous activity and network topology. ROI- and seed-based connectivity analyses further assessed age-related changes in distributed interactions.
Voxel-wise analyses revealed convergent age-related patterns across local activity and network organization. Increased ReHo and ALFF/fALFF with age were predominantly observed in memory- and control-related regions, including the hippocampus, cerebellum, thalamus, anterior and mid-cingulate cortex, and frontal association areas. In contrast, age-related decreases were mainly located in sensorimotor and default-mode regions, such as the middle temporal gyrus, precentral gyrus, and bilateral precuneus. DC analyses further indicated an age-related redistribution of network hubs, characterized by increased centrality in bilateral parietal cortices and reduced centrality in the right superior and middle frontal regions.
ROI-based analyses focusing on the hippocampus confirmed robust positive associations between age and hippocampal ReHo, ALFF, and fALFF, all surviving Bonferroni correction, highlighting consistent age-related modulation of local hippocampal activity. At the network level, seed-based connectivity analyses revealed age-related reductions in functional connectivity between the hippocampus and the right ventromedial prefrontal cortex, inferior temporal cortex, and left supplementary motor area. DC-guided connectivity analyses further demonstrated increased coupling between the left superior parietal cortex and visual regions, alongside decreased connectivity between the right superior frontal cortex and bilateral visual and dorsal cortical areas.
These findings indicate coordinated age-related reorganization of local neural activity and large-scale functional networks in middle-aged to older pilots, reflecting adaptive and selective changes in brain systems supporting memory, control, and visuospatial processing, with potential implications for cognitive aging research and aviation safety evaluation.

As we age, changes occur in all areas of the body, including the brain, and this affects cognitive functions. This natural process can be critical in professions that require high cognitive performance throughout a career, such as air traffic controllers (ATCOs). The problem is further amplified as some European countries intend to raise the retirement age, leading to an aging workforce. From a practical perspective, air traffic service providers (ATSPs) need staff capable to perform their work tasks safely, regardless of their chronological age. However, to date, there is little knowledge about actual and perceived aging effects, individual differences in cognitive decline, the compensating effect of experience, adequate performance assessment, and potential mitigation measures.

As such, this study examined the effects of aging in ATCOs in a Dutch and Swedish ATSP. First, three workshops were held in the Netherlands with older ATCOs (aged 50 or higher). The outcomes of these workshops formed the basis for the composition of a comprehensive cross-sectional survey. The survey was distributed to all ATCOs in both the Netherlands and Sweden, and consisted of questions about experience, work-home balance, health, workability, need for recovery, sleep and coping with irregular working hours, and cognitive functioning/decline.

In total, N = 12 Tower/Approach ATCOs participated in the workshops (2 female, 10 male). The majority of the participants said they had the idea that their job had become more complex and busy. With increasing age, they found less able to cope with this, for which they have sought possibilities to reduce their overall workload (either through working less hours or dropping management tasks). Several interviewees specifically expressed increasing problems with irregular working hours and on-call duties, and the desire to avoid night shifts. Participants also indicated a desire to have regular objective performance measurements and (yearly) discussions with management to monitor possible (fast) cognitive decline and well-being.

In order to validate the findings of the workshop and to study possible differences between age groups, a cross-sectional survey was distributed. In addition, to learn more about potential (cultural) differences between countries, this questionnaire was used in both the Netherlands and Sweden.

"Younger generations do not like to follow rules", "Older generations are slow and stubborn" and "All generations think and act in the same way".
These remarks are heard when people try to understand why organisational procedures are not complied with. Because compliance with Standard Operating Procedures (SOPs) is critical to aviation safety, it is essential to understand whether (and how) different groups may vary in how consistently they apply and act in accordance with these SOPs.
Following literature, generations are described as socio-historically shaped birth cohorts (Woodward et al., 2015), but cohort labels and boundaries vary, with the extent to which they effectively explain differences between groups being debated (Korteling et al., 2019; Rudolph et al., 2021). Moreover, claims about "younger" versus "older" workers' rule-following and procedure compliance recur over time but lack consistent empirical support and are often entangled with contextual influences (Rauvola et al., 2019; Rudolph & Zacher, 2016).
So, how should aviation organisations manage and support SOP compliance when a large share of the pilot workforce consists of younger pilots?
The present study examines how organisations should deal with a large group of younger pilots in relation to SOP compliance, and whether differences in procedural adherence can be observed between age cohorts within the pilot population.
Using a mixed-methods design, we first distributed a questionnaire among pilots (N = 54) to capture demographics and self-reported SOP-related practices. Based on these findings, we conducted semi-structured interviews with nine pilots to explore recurring themes in how new and revised procedures are interpreted and implemented daily operations.
Results indicate that (1) no evidence was found that pilots from younger generation cohorts deviate more from new or revised procedures than other cohorts and (2) when deviations occur, they are mainly linked to routine/old habits, perceived (often self-imposed) operational pressure, or occasional safety-driven adaptations. These results point towards operational conditions rather than age cohorts as the primary lever for improvement.
In the end, it's not the year on the birth certificate that matters, but pressure, routine, organisational factors and context in the cockpit.

The European Union Aviation Safety Agency (EASA) aims to enhance air safety through guidelines and regulations addressing the professional aptitude, selection, and training of pilots. In this context, pilot selection and training seek to identify individuals with the potential to acquire and reliably perform the complex cognitive, perceptual, and psycho-motor skills required for safe flight operations. Traditionally, pilot performance during training is evaluated primarily through instructor-based assessments. While expert judgements are indispensable, such evaluations may be influenced by subjectivity and often provide only limited insight into the detailed development of performance. Advances in simulator technology now enable the continuous recording of high-resolution flight data, creating new opportunities for objective and data-driven performance assessment. Automated monitoring systems can capture key performance dimensions relevant to pilot competence, including workload management, situational awareness, and psycho-motor control.

This study investigates whether automated flight performance monitoring using AMO can identify patterns of performance development and workload sensitivity in simulator training. In particular, the study compares performance trajectories of ab-initio pilots and ready-entry pilots with prior flight experience.
Preliminary analyses (N=42) indicate that AMO-derived flight quality indicators demonstrate acceptable to good reliability.

The results reveal heterogeneous performance patterns: while some trainees show clear performance improvements, others exhibit performance decrements when additional tasks and increased workload are introduced. These findings suggest that automated monitoring can identify individual differences in learning trajectories and workload sensitivity. Integrating such objective indicators into training evaluation may complement instructor assessments and support more data-informed decisions in pilot selection and training management.

Assessing competencies is a complex and challenging task, particularly when it comes to identifying root cause. This exploratory two-part study concentrates on the assessment of non-technical competencies necessary for initial flight training with the Royal Netherlands Air and Space Force. The goal was to research how and to what extent both subjective and objective data can be used to support instructors in their assessment of non-technical skills, such as workload management.
In both parts of the study students were asked to fly an hour-long sortie in a VR-flight simulator encountering several special 'events' along the way. These events were created to induce behaviour related to the non-technical competencies. The students were rated by an assessor on several non-technical competencies and objective data from the simulator and eye-tracker were collected. After the flight the assessor debriefed the students and asked them to self-rate their flight on the same competencies. Simulator data was then processed to create objective variables and compared to subjective ratings. Results from the first study were presented during EAAP 34.
This second part study focuses on the non-technical competency's workload management and situational awareness. Results show, among other things, the difficulties of assessing non-technical competencies in a realistic setting, the advantages of using more surprise and unknown elements and underscore the necessity of variations in behaviour. Furthermore, subjective scores on workload management and situational awareness could both be predicted using objective data.
This study demonstrates the potential of combining subjective and objective data to support instructors in assessing non-technical competencies, such as workload management and situational awareness, in initial flight training.

This study was carried out in cooperation with the Royal Netherlands Air Force 131 squadron (initial flight training) and funded by the RNLASF centre for Innovation, AIR.

Prospective memory (PrMe) refers to the ability to remember to carry out intended actions after a delay. This function is essential for pilots who must resume suspended tasks under a high workload. We examined whether trait mindfulness predicts PrMe performance of airline pilots in two different studies. In the first study, which relied entirely on self-report measures, the mindfulness trait (both general mindfulness and piloting job-specific mindfulness) was positively associated with perceived PrMe performance, suggesting that mindfulness is a significant determinant of pilots' subjective evaluations of their prospective memory. In the second study, however, using objective behavioral measures during realistic flight with a Boeing 737 NG full-flight simulator, no significant relationship was found between trait mindfulness measures and PrMe performance.
This divergence raises an important methodological and conceptual question: Why does trait mindfulness predict perceived PrMe but not objectively measured operational performance? One possible explanation for this discrepancy concerns the well-documented misalignment between self-assessed psychological constructs and objective performance outcomes in high-stakes training environments. Research in performance calibration has shown that individuals' judgments about their own competence do not always accurately reflect actual behavioral ability, particularly under complex and dynamic task demands. In aviation contexts, pilots have been found to overestimate their performance when relying solely on subjective evaluations, suggesting that self-perceptions may capture confidence or perceived competence rather than observable task execution.
A second consideration relates to construct measurement. Self-report mindfulness scales primarily assess perceived attentional tendencies rather than context-sensitive attentional control during real-time operational demands. Such instruments may therefore lack sensitivity to fluctuations in cognitive control under high workload conditions. Additionally, social desirability may influence responses, as attentional capacity is widely regarded as a core professional competence in aviation. This may increase the likelihood of inflated self-ratings.
Taken together, these results do not invalidate the role of mindfulness in aviation performance, but rather underscore the importance of distinguishing perceived cognitive strengths from objectively demonstrated performance. Future research should examine whether mindfulness training produces measurable improvements in simulator-based cognitive outcomes and explore its potential contribution to other operational domains such as stress regulation, adaptive decision-making, and crew coordination.

This study explores how commercial airline pilots experience Situation Awareness (SA) in their daily operations. It aims to understand pilots' own definitions of SA, the most common factors that degrade it, the personal strategies they use to maintain it and the real-world challenges they face when intervening in a fellow pilot's loss of SA. While theoretical frameworks like Endsley's three-level model have shaped our understanding of SA, there is limited research capturing the subjective, in-depth experiences of line pilots. A mixed-methods approach was utilized. An anonymous, purpose-built survey was distributed to airline pilots. The survey combined quantitative items with qualitative open-ended questions. Pilots were asked to describe SA in their own words, recount specific instances of SA loss and explain what helps or hinders them from speaking up when they perceive a colleague's SA is compromised. Data were anonymized and thematically analyzed. Quantitative results (N= 160) indicate that "fatigue", "high workload" and "distractions" are the most frequently reported causes of SA loss. A majority of pilots also reported occasional difficulty in balancing automation monitoring with external scanning. Qualitative thematic analysis of pilot narratives revealed three core themes: 1) The Essence of SA: Pilots consistently described "good SA" not merely as perception but as a dynamic state of "being ahead of the aircraft" possessing a coherent mental model and maintaining a "overarching view" of the operation. 2) Pathways to SA Loss: Narratives of SA loss events frequently described moments of "tunnel vision" during high workload, confusion induced by unexpected automation behavior and the insidious onset of complacency during routine phases of flight. 3) The Intervention Paradox: Critically, while pilots acknowledged the importance of intervening, they identified significant barriers to action. This pilot-centered study provides rich, empirical insights into the real-world dynamics of SA. The findings confirm that SA is not just an individual cognitive trait but a collaborative team process shaped by communication, hierarchy and organizational culture. The identified barriers to intervention -particularly those related to rank and communication norms- highlight a critical gap between theoretical CRM ideals and cockpit reality.

Aviation systems are increasingly constrained not by technological capability, but by human cognitive limits under sustained and interacting task demands. While air traffic controllers and pilots routinely operate near peak workload, prevailing automation paradigms remain largely static responding to momentary system states rather than the accumulated cognitive burden experienced by human operators over time. As a result, fatigue, stress, and performance degradation are often detected only after safety margins have already eroded. Building on the construct of Cumulative Cognitive Complexity, previously introduced and empirically validated as a temporal extension of task complexity, this paper examines how cumulative task demand can be leveraged as a central organizing variable for dynamic human performance management in aviation. Cumulative cognitive complexity integrates task characteristics, time-on-task, and contextual system demands to explain non-linear changes in workload, fatigue, vigilance, and error risk across operational duties. Drawing on evidence from six real-time field studies involving more than 130 air traffic controllers, including psychometric measures and EEG-derived indicators of cognitive fatigue and workload, the paper demonstrates that Cumulative Cognitive Complexity provides predictive insight into human performance trajectories well before observable performance breakdown occurs. These findings highlight the limitations of static readiness assessments, fixed staffing models, and role-based automation concepts. The paper introduces STORMS (Socio-Technical Operational Resource Management System) and the Synapse Human Performance Calculation Engine as applied implementations of this approach. Together, they operationalize Cumulative Cognitive Complexity to support real-time, predictive allocation of human and technological resources across socio-technical aviation networks. We argue that Cumulative Cognitive Complexity offers a pragmatic and immediately deployable foundation for the use of AI and higher automation in aviation - especially at a Network level, not as substitutes for human expertise, but as adaptive regulators of cognitive load. By aligning automation with the dynamic limits of human performance, this approach provides a concrete pathway to enhanced safety, efficiency, and workforce sustainability.

Pilot mental workload emerged as a critical priority in aviation psychology during the 1970s and 1980s, following landmark accidents that revealed the limits of human attention under complex multitasking demands. Early human factors work focused on ergonomics; however, by the late 1980s, the field recognized the need for controlled, replicable laboratory paradigms to measure workload without requiring full-scale simulators or pilot participants.

The Multi-Attribute Task Battery (MATB), first introduced by Comstock and Arnegard (1992), is a microworld comprising four concurrent aviation-like subtasks. MATB represented a milestone in aviation psychology by standardizing task structure, event logging and difficulty manipulation. This development enabled systematic comparison across laboratories and contributed to fundamental research on automation, workload, attention management and performance under stress. Over the course of three decades, MATB underwent a series of transformations, first through MATB-II (Santiago-Espada et al., 2011), and subsequently to the open-source version OpenMATB (Cegarra et al., 2020). Each of these adaptations was driven by the shifting landscape of technological capabilities and the evolution of research priorities. However, existing implementations require local installation and IT-background, limiting sample diversity and constraining large‑scale, remote, or field research.

This contribution presents WebMATB, a fully browser‑based implementation that preserves the classic MATB task logic and interface while removing installation, platform and hardware barriers. WebMATB runs on standard laptops, tablets and smartphones, democratizing access to a tool that has historically required specific laboratory setups or IT-background. A second novelty is its scenario authoring workflow. While scenarios remain compatible with the standard text format to support continuity with existing protocols, WebMATB introduces a dedicated graphical scenario editor that allows experimenters with limited technical skills to design, modify and disseminate scenarios, lowering the entry barrier for study creation and promoting protocol standardization across sites.

WebMATB exemplifies how foundational paradigms in aviation psychology can be adapted to contemporary research ecosystems with cloud‑based data collection, multi‑site collaboration and physiological integration via Lab Streaming Layer while maintaining continuity with 35 years of MATB literature. By bridging this legacy with modern web technologies, WebMATB aims to broaden participation, support large‑N remote studies, and facilitate replication across the aviation psychology community.

The European Plan for Aviation Safety (EPAS 2026) identifies loss of tacit knowledge (SI-3008), training effectiveness gaps (SI-3011), and knowledge transfer challenges for new aviation personnel (SI-5033) as critical human factors issues amid workforce shortages and ageing pilots. As senior captains retire, digital YouTube channels by self-identified aviation mentors emerge as informal training ecosystems, potentially compensating for formal mentoring deficits. This study examines their creator content to assess safety promotion potential.
An archival netnography (Kozinets, 2015) of 10 high-reach YouTube channels (>50k subscribers; EASA-relevant A320/737 instructional content, 2023–2026) analyses ~100 videos (~50 hours of transcripts) and metadata. Multi-layer thematic analysis combines inductive coding (typologies of transmitted knowledge) with deductive overlays from CRM/non-technical skills frameworks (NOTECHS) and EPAS safety issues. A structured 5-point instructor rating rubric (accuracy/SOP, NTS/CRM, tacit transfer, pedagogy; target ICC>0.8) yields quantitative effectiveness scores (N=100 videos).
Preliminary analyses indicate emerging themes related to line‑operations heuristics, CRM debriefing practices, and the normalization or correction of procedural deviations, each of which can be aligned with current EPAS priority areas. The study's findings will clarify whether digital mentoring mechanisms strengthen or undermine safety culture, thereby providing regulators, airlines, and training organizations with evidence on how informal digital channels may be leveraged to support the preservation and transmission of tacit operational knowledge.

Aviation training environments demand more than technical expertise from instructors; they require strong interpersonal skills, emotional stability, leadership capacity, and the ability to manage performance under pressure. Recognizing this need, the Aviation Instructor Profiling Inventory (AIPI) was developed as a structured personality–competency assessment tool. The primary goal of AIPI is to provide a standardized, evidence-based framework for evaluating aviation instructor candidates in both cockpit and cabin training contexts.
The inventory assesses key domains including competency-based areas such as application of knowledge, communication, interaction with the trainees, instruction, management of the learning environment, as well as behavioural characteristics including clear and direct feedback, analytical thinking, flexibility and adaptability, situation awareness, empathy, persuasion and influencing skills, openness to different perspectives and feedback, fostering a sense of safety, leadership, self-respect and self-awareness, planning and organisation, openness to change, emotional resilience, and initiative-taking. Its development followed a multi-stage psychometric process. First, the draft format was reviewed by eight subject-matter experts to ensure content relevance and conceptual clarity. Following revisions, the instrument was implemented with 180 instructors. Afterwards, data were collected from 344 cockpit and 291 cabin candidates.
The assessment format integrates multiple item types to capture behavioural tendencies more comprehensively. For cockpit candidates, the inventory includes six scenario-based and eight short-form multiple-choice items; for cabin candidates, eight scenario-based and ten short-form items are used. Additionally, three (cockpit) and five (cabin) open-ended questions provide qualitative insights. A peer mentor scale, consisting of 35 items, was also incorporated to introduce multi-source evaluation and strengthen convergent validity evidence. The peer mentor scale comprises three sub-dimensions: guidance and mentoring, communication and openness to collaboration, and emotional regulation.
Item analyses, scoring algorithm refinements, and sub-dimension structuring were conducted based on statistical performance indicators and interpretability criteria. Internal consistency and validity analyses supported the psychometric robustness of the instrument. Furthermore, the inventory was calibrated with behavioural interviews.
Overall, AIPI represents a shift from solely interview-based evaluations toward a more systematic and data-driven profiling approach. Ongoing work includes continuous item revision, expansion of the question pool, periodic recalibration, and longitudinal validation to examine predictive utility in instructor performance outcomes.

Touchscreen displays are becoming more prevalent in large aircraft; they enable designers to reduce display clutter, provide advantageous information structure, and save weight. Touch as a primary modality of interaction; however, raises concerns for time on task, performance accuracy, workload, fatigue, undue concentration, error rates, and usability in turbulence. Level D flight training simulators are not designed for Human Factors (HF) data collection and don't have the necessary representativeness or controllability of turbulence. Real aircraft flights are expensive, and pilots are unpredictably exposed to varying degrees of turbulence which eliminates repeatability and the ability to meaningfully apply statistical methods for data comparison and analysis. An HF evaluation facility was purpose-built by Honeywell Aerospace in Brno, CZ; capable of simulating representative turbulences and using modular physical architecture (e.g., flight deck geometry and installed equipment) to allow customization to most aircraft types. This facility was designed specifically for HF data collection by supporting a six-person evaluation team using multiple cameras, custom audio communications and evaluator stations, a proprietary turbulence model, recording equipment, and separate observer room. Full control over the simulated turbulence allows collection of data which shows all participants experience the same conditions for the same tasks at the same time, all while pilots perceive the turbulence levels as random, with no learnable patterns. In this facility, Honeywell Aerospace conducted a simulated flight from Paris to Frankfurt with seven pilots, simulated Air Traffic Control and representative pilot tasks to collect data used to alleviate HF concerns of implementing touch as a primary interaction modality for large airplanes. The successful outcome of this evaluation demonstrated the utility of the facility and data collection methods. To validate physical representativeness (i.e., flight deck geometry and hand stabilization), functional representativeness (i.e., hardware and software), and operational representativeness (e.g., turbulence profile and pilot tasks) a rating scale was developed, and geometry and inertial data was compared (platform versus real aircraft). A large European aircraft manufacturer developing a new aircraft and the responsible certification Authority accepted this new facility for the collection of HF data to contribute to compliance demonstration with applicable airworthiness regulations.

The safety climate of an organisation provides a critical snapshot of how effectively safety goals are prioritised, communicated and enacted in day-to-day operations. While many existing instruments focus primarily on employees' safety attitudes or perceptions, these approaches may not fully capture how safety is actually practised across organisational levels. This presentation introduces the development of an innovative safety climate tool that moves beyond attitudinal measurement to directly assess observable safety-related behaviours and practices among leadership, middle management and frontline staff. By focusing on what people do, in addition to how they feel, the tool offers a more behaviourally grounded and operationally meaningful picture of organisational safety. The instrument was developed through a rigorous review of contemporary safety culture and safety climate literature, integrating established psychological theories with applied insights from safety practitioners. Particular attention was given to multi-level influences on safety, recognising that leadership actions, supervisory practices and individual behaviours interact to shape safety outcomes. Items were carefully constructed to reflect real-world safety practices, including communication, accountability, learning from incidents, resource allocation and reinforcement of standards. Psychometric evaluation demonstrates strong internal reliability and a coherent factor structure consistent with the underlying theoretical model. The results support the stability and construct clarity of the tool across organisational groups. Importantly, the assessment is designed not only as a diagnostic measure but also as a practical change instrument. Findings are translated into actionable insights, enabling organisations to identify perception gaps across job levels, pinpoint behavioural inconsistencies and prioritise targeted interventions.

Commercial aviation operates within a constant tension between production and protection. While safety must never be subordinated to economic objectives, airlines continuously find themselves balancing operational efficiency and risk control. Modern Safety Management Systems, as required under ICAO Annex 19 and EU regulations, formalize hazard identification, risk assessment, and mitigation measures. Yet, the most valuable safety intelligence does not emerge from mechanical mathematical processes, but from operational personnel.
Occurrence reports from frontline professionals represent the primary source of actionable safety data. However, proactive reporting is based on psychological safety and mutual trust. Without a downright Just Culture framework, staff might withhold safety-sensitive information, leaving critical conditions undetected. Using James Reason's famous theoretical framework, this presentation emphasizes the fact that a single root cause is hardly ever the main trigger of accidents. Moreover, accumulating underlying threats, embedded within organizational processes, hardware design, inexplicit procedures and/or human behavior have been identified as the essential contributing factors in accident investigation.
Data analysis (esp. flight data monitoring, reporting, rostering) and well-structured investigations (acc. to ICAO Annex 13, tailored to operator-specific needs) transform individual events into valuable systemic learning opportunities. Applying the Human Factors Analysis and Classification System (HFACS) by Douglas Wiegmann and Scott Shappell (2000), the process is centered around preconditions for unsafe acts, potential unsafe supervision, and organizational influence, rather than the unsafe act itself. The aim of this analysis is to uncover the chain of events to generate evidence-based safety recommendations. Investigations performed within a mutually confident Just Culture environment aim to uncover potential system vulnerabilities rather than holding someone accountable. Interviews focusing on human factors mostly reveal a wide range of information on cognitive workload, operational pressure, communication dynamics, and decision-making constraints that are hardly ever visible in raw data analysis only.
Selected cases and events demonstrate how an effective reporting culture together with human factors-driven investigations push the shift from reactive learning toward a proactive safety culture. Thus, the foundation of aviation safety is based upon fostering an environment in which staff can speak out openly, and where systems are willing to learn and improve continuously.

Just Culture frameworks underpin aviation Safety Management Systems, yet their operational application frequently exposes tensions between learning and accountability. Traditional culpability decision trees have been criticised for encouraging hindsight bias, oversimplified categorisation, and premature escalation to disciplinary processes. In practice, these weaknesses can erode reporting trust and undermine psychological safety, both critical to effective safety culture. This paper describes the applied development of a revised Just Culture Case Assessment framework designed for operational aviation environments. Drawing on Reason's error typology, Dekker's concept of local rationality, and Marx's behavioural accountability model, the framework was iteratively refined to address recurring implementation problems observed in airline, maintenance, and airport operations. These included outcome-driven reinterpretation of behaviour, automatic escalation of risk-aware decisions to disciplinary review, and insufficient recognition of how operational pressures shape judgement and risk perception. The framework separates safety investigation from behavioural assessment and introduces an independent assessment function to mitigate bias and role conflict. It requires action-level classification, prohibits retrospective reclassification based on repetition, and mandates examination of organisational signals where mission-driven risk acceptance occurs. From an applied aviation psychology perspective, the model integrates cognitive bias mitigation, risk perception under pressure, and organisational influence on decision-making. Early implementation indicates improved consistency, reduced inappropriate disciplinary escalation, and strengthened trust in reporting systems while preserving a defensible threshold for reckless conduct.

The evolution of modern aviation has shifted focus of safety from mechanical reliability to complex interplay of human factors, specifically within the framework of Crew Resource Management (CRM). While automation has significantly reduced manual flight errors, it has introduced new cognitive challenges related to trust calibration. This research investigates the role of trust as a multi-dimensional construct in aviation incidents and accidents by analyzing a vast corpus of open-access investigation reports. The study employs a structured scoring methodology to evaluate two distinct yet interconnected dimensions: 'Trust the Machine'referring to the pilot's reliance on and interaction with automated flight systems—and 'Trust the Cockpit Team'—referring to interpersonal dynamics, psychological safety, and communication efficacy between flight deck members.The methodology utilizes a systematic content analysis of narrative data from agencies such as the NTSB and SKYbrary, spanning the last decade of reported incidents. Each report is evaluated against a standardized rubric that scores indicators of over-reliance, distrust, and miscalibration in both dimensions. By mapping these scores against established CRM markers like situational awareness, assertiveness, and decision-making, the study identifies critical failure points where trust becomes a liability rather than an asset. Preliminary findings suggest a dual-risk landscape: on one hand, 'automation complacency' leads to a degradation of manual monitoring skills when trust in the machine is absolute; on the other hand, a lack of 'interpersonal trust' or high-power distance within the cockpit inhibits the vital cross-check functions necessary for error detection.The analysis reveals that many accidents are characterized by a 'trust-transparency gap' where pilots either over-estimate the system's logic or under-communicate their doubts to their colleagues due to a perceived lack of psychological safety. The findings to be presented at this conference will detail the statistical distribution of these trust scores across various accident categories, illustrating how miscalibrated trust frequently acts as final link in error chain. Furthermore, the presentation will provide evidence-based recommendations for integrating trust-management protocols into future CRM training modules. By quantifying these intangible human elements, research offers a novel framework for enhancing flight deck resilience in an increasingly automated environment, bridging gap between technical system design and human psychological realities.

1. Operational experience is widely regarded as a cornerstone of aviation safety. However, evidence from flight operations indicates that under certain conditions, high experience combined with authority gradients, expectation bias, and routine exposure can unintentionally introduce latent safety risks. This phenomenon - Captain-induced error remains under-recognized within traditional Safety Management Systems (SMS), which often prioritise low-experience risk while overlooking experience-related cognitive bias.

2. This paper/presentation explores captain-induced errors through the lens of key safety priorities: runway safety, approach stability, and ATM–flight deck interaction. Using de-identified operational cases, the session examines how experienced captains may unintentionally normalise unstable approaches, accept marginal runway conditions, or discount ATC inputs due to expectation bias and past success. These behaviours are analysed using Threat and Error Management (TEM) and HFACS, supported by operational data indicators such as unstable approach trends, runway excursion precursors, and deviations in standard crew-ATC communication patterns.

3. The paper/presentation further examines how data-enabled safety tools and AI-supported analytics can assist organisations in identifying experience-related risk precursors. Examples include monitoring approach stability exceedances, Captain-First Officer intervention rates, and communication asymmetry during high-workload phases. Emphasis is placed on human-in-the-loop design, explainability, and just culture safeguards to ensure that data and AI enhance rather than undermine professional judgement.

4. Designed as an interactive workshop, the presentation of this paper will combine short expert inputs, facilitated discussion, and scenario-based audience engagement. Participants will collaboratively identify practical SMS interventions, CRM enhancements, and data-driven feedback mechanisms applicable to European operational contexts. The outcome is a pragmatic framework for integrating human performance insight with data-enabled safety strategies, strengthening runway safety, approach stability, and ATM integration without eroding the value of experience.

Unintentional actions on cockpit controls, commonly referred to as cockpit control substitution errors, represent a persistent safety concern in commercial aviation. These errors occur when pilots inadvertently operate an incorrect control or execute an action in an unintended sequence, often despite high levels of experience and familiarity with aircraft systems. Rooted in human cognitive limitations rather than deficiencies in knowledge or skill, such errors provide a valuable lens through which to examine human error mechanisms and prevention strategies.
This mixed-methods study investigated the prevalence, contributing factors, and mitigation strategies associated with cockpit control substitution errors in modern commercial flight operations. Semi-structured interviews were conducted with fourteen commercial pilots, followed by a quantitative survey completed by twenty-six pilots. Qualitative data were analyzed using thematic analysis, while quantitative data were examined using non-parametric statistical methods to explore relationships between error recency and pilot-, aircraft-, and system-related variables.
Findings indicate that cockpit control substitution errors are relatively common and occur across experience levels. Key contributing factors include automatism and habitual action, muscle memory, cockpit ergonomics, high cognitive workload, stress, and limited training emphasis on attentional control during routine operations. Quantitative results revealed statistically significant associations between error recency and pilot role, experience level, aircraft type, system design, and perceived ergonomic characteristics.
The results indicate that complete elimination of such errors is unrealistic; however, their likelihood and operational impact can be meaningfully reduced. An integrated prevention approach is therefore proposed, combining procedural discipline, targeted training interventions, and design-oriented defenses. Recommended measures include enhanced training focused on deliberate action and attentional control, reinforcement of standard operating procedures that encourage verification and pauses before action, strengthened crew cross-monitoring practices, and continued human-centered improvements in cockpit design and system feedback.
By providing contemporary empirical evidence and practitioner-informed insights, this research contributes to aviation psychology by advancing understanding of human error mechanisms and supporting the development of practical, system-level prevention strategies to enhance flight safety in increasingly complex flight deck environments.

Error reporting and disclosure are central elements of aviation safety systems and play an important role in learning processes during flight training. In aviation contexts, acknowledging and discussing errors supports both individual learning and the continuous improvement of safety practices. At the same time, the willingness to report or disclose mistakes may depend on how psychologically safe individuals feel within the training environment. While psychological safety has received increasing attention in organisational research, little is known about how ab-initio pilots perceive and experience psychological safety and error disclosure during flight training.Understanding these perceptions is important because attitudes toward mistakes during training may influence future safety behaviours in aviation operations.
This study explores how ab-initio pilots experience and perceive psychological safety in relation to error reporting and error disclosure during flight training, and how these perceptions relate to their decisions about whether and how mistakes are reported or disclosed.
This research adopts an exploratory qualitative design using reflexive thematic analysis and is currently in the interview phase. Semi-structured one to one interviews are being conducted with 12 ab-initio pilots who are currently in the third and fourth year of an undergraduate pilot training programme. All participants hold a Private Pilot License and are continuing their training toward an Airline Transport Pilot License.
The interviews explore participants' experiences and perceptions regarding making mistakes during flight training and psychological safety within the training environment. An example interview question is: 'How comfortable or safe do you feel discussing mistakes in your training environment?". The interviews are being analysed in NVivo15 using inductive reflexive thematic analysis, allowing themes to emerge from participants' accounts.
Preliminary insights suggest several potential themes, including Selective Disclosure of Errors, Instructor–Student Power Distance, and the Judgment vs. Learning Dilemma.
By examining how ab-initio pilots perceive psychological safety when dealing with mistakes during flight training, the study highlights the role psychological safety plays in whether and how errors are reported or disclosed. These insights may help inform efforts to support psychological safety in flight training environments and aviation organisations, particularly regarding how future pilots acknowledge and discuss mistakes in safety critical contexts.

Precise identification and description of individual human error remain foundational challenges in aviation psychology and human factors research. Despite extensive taxonomic development, inconsistencies persist in how errors are characterised, how non-compliance is distinguished from unintentional failure, and how individual performance is analytically linked to system interfaces. This paper presents a SHELL grounded human error classification model designed to strengthen both the theoretical clarity and operational reliability of individual error analysis. The innovation of the model lies in its explicit separation between observable deviation, point of manifestation within the human system interface, and internal psychological process. By structuring classification in this way, the framework reduces hindsight bias, limits premature attribution of causation, and promotes disciplined inference regarding cognitive, perceptual, decisional, monitoring, learning, and intentional factors. This repositioning of psychological constructs as evidence based layers of analysis, rather than default explanatory starting points, represents a methodological contribution to applied cognitive psychology in safety critical environments. Operationally, the model embeds individual error identification directly within the SHELL architecture, requiring analysts to specify where an error manifested across hardware, software, environmental, interpersonal, or internal domains. This strengthens traceability between individual performance and system design characteristics, supporting clearer translation from incident data to intervention. The inclusion of monitoring and learning failures extends traditional classifications by addressing performance drift and recurrent error patterns that are often under specified in existing approaches. The framework contributes to aviation psychology by enhancing conceptual precision in the treatment of individual error, and to human factors research by improving reliability, comparability, and analytical transparency in occurrence investigation. By aligning psychological theory with structured interface analysis, the model offers a practical yet theoretically coherent advancement in the classification of human error within complex aviation systems.

Aviation has long prided itself on being a global industry that connects people from all corners of the world. Consequently, individuals from diverse backgrounds are professionally engaged in aviation, prompting increased research interest in diversity within the sector. This paper presents a narrative review of the past thirty years of research on diversity in aviation and safety. The scope of this review included peer-reviewed articles across several databases published after 1990 with prime search terms "diversity", "aviation" and "safety" supplemented by "nationality", "gay", "LGBTQ", "ethnicity" and "culture".

The most frequently studied groups were women in aviation, people of color, individuals of diverse nationalities and cultures, and diversity in educational and professional backgrounds. In contrast, very little to no research has addressed disability, the LGBTQ+ community, or emerging diversity categories such as neurodiversity. The literature finds challenges such as biases at hiring, training, career opportunities, communication, financial support, cultural and social norms, and lack of role models to constitute a significant barrier to inclusion in aviation.

The review also finds a growing fatigue with diversity programmes and perceptions of performative affirmation initiatives. The overuse of role models, dissonance between corporate imagery and perceived organisational culture, and the presentation of statistics without sufficient contextualisation have all contributed to perceptions of "box-ticking" approaches to DEI leaving feelings of unfairness in both minority and majority groups.

The findings of this research call for future research on: (a) tailored initial- and continuation- training programmes in aviation that take into account local cultural context and social norms; (b) the development of interpersonal and cross-cultural competencies among aviation personnel as enablers of safety across borders; and (c) sustained efforts not only to attract more women and individuals from underrepresented groups to aviation, but also to reduce barriers to their long-term career progression and success.

Mental health remains stigmatized in aviation, with 18.5% percent of pilots admitting to intentionally avoiding seeking assistance with mental health problems. Because of fear of potential loss of license (either temporarily or permanently) and the consequential loss of income and identity, pilots will often engage in very specific behaviors in which they simultaneously avoid seeking formal, documented treatment while preferring to hide symptoms and seek informal (and sometimes unproven) assistance.
In this study of 133 Icelandic pilots, participants were asked about the likelihood of seeking assistance with mental health-related problems and then specifically asked what could be done to make them more likely to seek mental health assistance if needed.
When given specific options of improved help-seeking 41% of pilots said that they would be more likely to seek assistance if the psychologist had an aviation-specific credentials (e.g. EAAP accredited aviation psychologists); 49% of pilots stated that a provision of union-endorsed psychologists would make them more likely to seek assistance; 53% believed that an implementation of a SafeHaven approach (as currently being implemented by New Zealand and Australia) where an AME can coordinate treatment with a psychologist without reporting it to the local CAA and invalidating the pilot's license, would be beneficial and lead to more help-seeking from pilots suffering from mental health problems; and 56% of pilots felt that by having the option of being directly referred by their local Peer Support Program would make them more likely to seek assistance.
This proposed presentation will discuss these results in more detail and provide thematic and content analyses of open-ended questions about what could be done to increase mental health-seeking behaviors among pilots.

While the Job Demands-Resources (JD-R) model is a well-established framework, its application in elite military aviation often overlooks the nuanced interplay between different social resources. This study examines how role-related demands—overload, conflict, and ambiguity—affect burnout and job satisfaction among active-duty Republic of Korea Air Force pilots. The sample (N=48) predominantly consists of fighter pilots (79.2%), representing a high-security, high-performance mission environment. Data collection was conducted via a secure military intranet, ensuring high integrity and participant anonymity.

Moderated regression analyses revealed that role ambiguity was the most powerful predictor of burnout, particularly cynicism and reduced professional efficacy. Crucially, the results highlight a structural hierarchy of resources: leader support effectively buffered the negative effects of job demands, whereas peer support showed a context-dependent effect and, in some instances, exacerbated stress responses when leadership was absent. This paradoxical finding suggests that in hierarchical aviation cultures, leader-led role definition is the primary driver of psychological resilience. These findings offer actionable insights for sustaining operational readiness. As aviation transitions toward Human-AI teaming where the boundaries of human responsibility are expected to become increasingly complex, proactive management of role clarity will be paramount to ensuring mission safety.

Objective: This presentation provides an overview of the progressive development of a support concept that addresses the needs of pilots with mental health conditions while ensuring flight safety, public safety, and compliance with EU aviation regulations.
Methods and results: First, a historical review outlines major stages in the development of a group of professional support programmes for pilots in Germany and the motivations underlying these changes. Second, the presentation highlights key differences between classical Pilot Peer Support (PPS) approaches (e.g., the EPPSI Guide) and a professional clinical-psychological concept integrating both support and monitoring components, including the role of trained pilot peers (PPSC). Standards for structural, procedural, and outcome quality in the assessment of fitness to fly and in the selection of tailored psychotherapeutic interventions are presented. Finally, programme benefits for airlines, pilots, and aviation safety are described, including significant cost savings for airlines and positive treatment outcomes based on approximately 600 cases.
Conclusions: Key elements of an effective treatment and monitoring system for pilots with mental health conditions include long-term, trust-based collaboration between the programme supervisor, psychotherapists, hospitals, peers, aviation medical examiners (AMEs), and the German Federal Aviation Office (LBA), as well as the integration of support and oversight and the involvement of trained and supervised peers.

The purpose of this work is to design a part of the instruction of cabin crew based on the normative education of Spanish secondary marketing to apply to crew resource management instruction. The title of the didactic unit is 'What can I do when facing a conflictive passenger?' The didactic objetive is to learn how to handle a disruptive passenger and do so without suffering emotional dissonance. It consist in ten sessions of one hour of duration and includes practical exercises in each of the sessions, within the framework of problem-based methodology and flipped classroom This didactic planning follows the method of Spanish formal secondary education. For this reason, it allows for rigorous and objective evaluation and grading by teachers during their interventions. The contents were obtained from aeronautical regulations and other intervention models and they constitute the students' educational competencies. The content was obtained from aeronautical regulations and various training programs from different authors and constitutes the students' educational competencies. This diseny is based in previous investigations. It has been applied in instruction for cabin crew and it has been evaluated as excellent by experts supervisors teachers and as utilitable in their duties by cabin crew. So, this training program has internal and empirical validity and the results can replicated.

Startle and surprise are known to potentially incapacitate professionals who respond to emergency situations. In recent years, research has focused on developing strategies to help pilots mitigate the negative effects of startle and surprise. However, no such research exists for cabin crew, despite their important safety role. The objective of our two-phase study was to first obtain insight into the prevalence and impact of startle and surprise in cabin crew, and second to design and evaluate a startle management method designed specifically for cabin crew.
We performed a survey among 348 European cabin crew. It showed that 79.3% of cabin crew have experienced startle and/or surprise in a wide range of events, with medical incidents as the most reported. High levels of perceived stress were correlated with reported performance impairments and lasting anxiety. Perceived control over the situation was correlated with lower perceived stress score and lower long-term anxiety. This indicates that startle and surprise can have a significant impact on cabin crew, and that cabin crew would likely benefit from startle management training that instils a sense of control.
In the second phase, we evaluated a 4-step startle management method through interviews, a focus group and practical testing in a fire-fighting exercise in a fire-trainer. After performing the scenario, cabin crew rated whether the method had positive or negative effects, was difficult or easy to use, and perceived stress before and after applying the method. Participants rated method effectiveness and usability significantly more positively than neutral. Perceived stress decreased significantly after, compared to before, applying the method. Participants provided several suggestions for improving method content, and for approaches to integrate the method more effectively into existing training. The outcomes of this study provide direction on the design and use of self-management methods that can help cabin crew manage startle and surprise, ultimately improving flight safety.

Aircraft escape slides are safety features intended to be deployed when there is a need to rapidly evacuate an aircraft. Personnel who interact with aircraft doors include cabin crew, maintainers, pilots, caterers, gate agents and others.
Slides are designed for ease of activation during an emergency, yet this also means that inadvertent actions by personnel will sometimes cause slides to be deployed in non-emergency situations. Inadvertent slide deployments (ISDs) can be dangerous to those in the vicinity, are a source of schedule disruptions, and impose significant costs as slides are reinstalled and repacked.
The United States National Aeronautics and Space Administration (NASA) Engineering and Safety Center was approached by several airlines that were frustrated by a continuing and persistent incidence of ISDs, despite interventions intended to address the problem.
A study of ISD events revealed some of factors associated with these events. We found marked differences between ISDs that occurred during aircraft operations, and those that were triggered by the actions of maintenance personnel. Nevertheless, there were commonalities between ISDs, regardless of the personnel involved. These include the risks of automatic behavior that emerges when a task is performed routinely, design features of door and overwing exits, the impact of interruptions and non-normal situations on familiar tasks, and potential gaps between "work as imagined" and "work as done".

As passenger numbers continue to rise and cabin crew workloads intensify, crews must be prepared to respond to an almost limitless range of scenarios at 37,000 feet. The ability to adapt under pressure – drawing on the right knowledge, attitude and skills - is critical for crew, where a single decision may determine outcomes in an emergency.
CBTA has been increasingly adopted across the industry, helping to nurture these adaptive behaviours. However, even the most robust training systems are influenced by individual differences, which can affect training success.
The selection process is usually the first 'Assessment' in CBTA, but how does this typically inform subsequent training, if at all? In pilots, the selection process is highly standardised and led by the core competencies for the role, helping to predict who is likely to be successful prior to starting training. Despite cabin crew also playing a safety critical role, similar industry standards or guidelines do not exist, and recruitment can sometimes be cursory, focussing on communication skills and appearance, but failing to look at the whole picture. Yes, customer service and professionalism are important – but will these traits alone ensure the selected crew are ready to react and utilise their training when faced with the unexpected?
During this session, we will share our survey findings on the current state of cabin crew selection and explore how we have used data from across the industry to design psychological assessments for cabin crew that address real-world selection needs and challenges. By identifying core competencies prior to training, selection can better predict training success, reduce attrition, inform targeted development, and strengthen confidence in cabin crew competence. We will discuss a case study from one of our launch clients, sharing the findings and outcomes of these tailored cabin crew assessments.

Aviation medicine recognises subtle incapacitation, referring to situations in which flight crew remain conscious and operational yet experience degraded cognitive performance. Traditionally examined through fatigue research, physiological monitoring, and accident investigation, this concept highlights the difficulty of detecting transient reductions in attention, processing speed, and situational awareness before they lead to operational error. As aviation systems increasingly explore physiological sensing and neurocognitive screening to monitor pilot state, questions remain about how subtle cognitive changes are recognised and managed in practice.
Parallel discussions in healthcare research have examined experiences often described as "brain fog", characterised by fluctuating attention, slowed thinking, forgetfulness, communication difficulties and unpredictable changes in cognitive clarity. Importantly, this literature emphasises not only cognitive symptoms but also how individuals interpret, describe, and adapt to these changes within everyday work and social contexts. Themes such as the invisibility of impairment, stigma surrounding disclosure, and the development of compensatory strategies highlight the social and organisational dimensions of fluctuating cognition. Discussions emerging from post-covid "brain fog" communities and literature further illustrate how individuals experiencing uncertain cognitive changes often rely on informal or peer-based support when symptoms are poorly recognised within formal systems, highlighting the importance of supportive cultures for recognising and managing emerging cognitive difficulties.
This paper adopts a speculative, cross-domain perspective to explore what aviation psychology might learn from these insights. While aviation research has largely approached subtle incapacitation through physiological and performance measurement, qualitative discussions in healthcare contexts illustrate how people themselves recognise, interpret, and manage cognitive variability in real-world settings. Such perspectives may complement existing monitoring approaches by illuminating how professionals recognise emerging cognitive changes, how organisational cultures shape disclosure, and how adaptive strategies are developed to maintain safe performance.
By synthesising insights from healthcare and aviation literature, this paper argues for a broader human-centred framework for understanding fluctuating cognition in operational environments. Recognising how cognitive variability is experienced and managed may support more nuanced approaches to safety culture, pilot state awareness, and the design of resilient human–technology systems.

Safe maintenance of UK military aircraft requires identifying and mitigating risks to a residual level which can be claimed As Low As Reasonably Practicable (ALARP) to enable ownership by military Duty Holders. The ALARP claim, argument and evidence is detailed in a Safety Case and supported using BowTie barrier-based risk management tools.
A key challenge is assuring the effectiveness of BowTie barriers which rely on human performance. This challenge is twofold: firstly, the barrier will be allocated an effectiveness level based upon qualified assumption which may be inaccurate. Secondly, barrier performance may differ from the assumed effectiveness level due to human performance variability. Therefore, the military Duty Holder may be holding a level of residual risk which is more variable than described in the Safety Case. This paper describes practically addressing this challenge by applying Human Factors (HF) analysis to increase the accuracy of the military Duty Holder ALARP position. The authors describe a case study in which maintenance personnel reported an increased frequency of road vehicle incursions into the aircraft manoeuvring area during aircraft ground towing operations. HF analysis of the reports indicated that vehicle drivers are not adhering to 'stop' lights. Task observation of aircraft towing identified that personnel were activating 'stop' lights after drivers had entered the manoeuvring area. Eye-tracking data from a vehicle driver perspective showed that the 'stop' lights were positioned outside the vehicle driver's primary field of view, and were obscured by vehicle structure. Moreover, eye-tracking data indicated that a low sun position reduced the salience of 'stop' lights. The HF analysis identified that traffic control barriers were less effective than assumed in the analysis supporting the Safety Case, while the human performance barrier of aircraft tow team vigilance was more effective than assumed. In addition to recommended improvements to traffic control barriers, this paper discusses how HF analysis can enhance the data used during qualified assumption of barrier effectiveness. The case study demonstrated the value of using HF analysis to assure risk barrier effectiveness to increase the accuracy of the Safety Case describing the ALARP risk level owned by the military Duty Holder.

This paper will provide an historical perspective on how our insight in the "somatogravic illusion" progressed over the years. The vestibular illusion involves an incorrect perception of linear accelerations in-flight, which are perceived as an attitude change, rather than an acceleration. The spatial disorientation illusion was described for the first time as a flight safety risk about one hundred years ago, but for a long time after that, accidents related to the illusion were attributed to "pilot error". In the era of fast jets, the illusion was identified as a causal factor for many controlled-flight-into-terrain (CFIT) accidents, where the unfortunate pilots made nose-down inputs in response to an acceleration-induced false pitch-up sensation, for example during a catapult launch. We will review some key psychophysical experiments from those years that improved our understanding of the illusion, as well as training approaches to familiarize pilots with the effects of the illusion. For a long time, the illusion was associated with high-performance aircraft. In the 2000's the awareness grew that the illusion can also occur in slower moving transport aircraft, especially during the go-around phase where the thrust-to-weight ratio is high. A French safety study into a number of such accidents inspired the European Union Aviation Safety Agency (EASA) to recommend airlines to perform go-arounds with reduced thrust. The rationale was that, apart from reducing time pressure in the cockpit, this would minimize the illusion. However, detailed analysis of go-around accidents suggests that, although these CFIT events show persistent nose-down inputs by the pilot, the inputs do not seem to be triggered by thrust-induced acceleration. Remarkably, model simulations suggest that pilots will even experience a more pronounced somatogravic illusion during go-arounds with reduced thrust, due to downward pitch adjustments needed to maintain air speed. It thus seems that one hundred years after the first documentation of the somatogravic illusion we still do not fully understand the way it may present itself to pilots.

Operational domains such as aviation, military operations, and healthcare are commonly described as high-risk environments (HREs). These contexts involve complex socio-technical systems in which human performance, organizational processes, and technological infrastructures interact under conditions where operational errors may have severe consequences. In such settings, human factors, safety culture, and effective error management are critical for maintaining safe and resilient operations. Personnel are often exposed to elevated cognitive, emotional, and psychological demands compared to conventional work settings. Despite widespread use of the term, a consistent and empirically grounded definition of high-risk environments from a work and organizational science perspective is lacking. This conceptual ambiguity complicates systematic identification of HREs and limits the development of standardized risk assessment, safety management, and occupational health interventions.

This study addresses the gap by systematically examining the concept of high-risk environments and identifying their defining characteristics. A systematic literature review was conducted to collect and analyse research on structural, operational, and human-factor-related features of HREs. From an initial pool of 13,556 publications, 121 studies were selected for detailed analysis. The extracted characteristics were synthesized into an integrated conceptual framework illustrating dynamic interactions and interdependencies between organizational, technological, and human factors shaping risk exposure in these environments.
Based on this theoretical framework, a questionnaire evaluated its practical relevance and applicability. The survey was completed by 43 professionals working in occupational domains typically associated with HREs. In addition, expert interviews were conducted to further validate and refine the framework. Combined insights from literature analysis, practitioner feedback, and expert evaluation were used to refine the framework and derive a definition of high-risk environments from an occupational science perspective.

The resulting framework provides a structured set of characteristics enabling systematic categorization and assessment of work environments according to risk profile. Theoretically, the study contributes conceptual clarity by offering a research-based definition of HREs and establishing a shared foundation for future human factors research. Practically, it offers safety-critical organizations like those in aviation a standardized basis for identifying risk-intensive contexts, supporting consistent risk assessment, informed resource allocation, and the development of targeted safety and protective measures.

Understanding how acute stress affects (military) pilot performance is critical for mission success and aviation safety. Within the Royal Netherlands Air and Space Force (RNLASF) we are experimenting with methods of stress inducement to train novice military pilots. Inducing stress within realistic flight environments is challenging, especially in novice pilots. Traditional aviation stressors such as simulated emergencies may be ineffective in this population, as limited technical knowledge can lead to confusion or passive responding. Laboratory stress paradigms such as the Cold Pressor Test, Trier Social Stress Test, and Maastricht Acute Stress Test (MAST) are well validated but rarely integrated into flight settings, limiting ecological validity. This study examined the feasibility and performance effects of inducing acute stress during flight in a virtual reality (VR) military training cockpit.

In a randomized controlled between-groups design, twenty military participants, without prior military flight experience, were assigned to either a stress or control condition. The MAST was adapted for in-flight implementation within the VR cockpit, enabling controlled stress induction while preserving ecological task realism. Following standardized training and a familiarization flight, participants completed an experimental flight. During this flight they executed four standardized maneuvers (straight-and-level flight, 90° turn, 360° turn, and speed change). Technical performance was operationalized as mean absolute altitude deviation from the assigned altitude (5000 ft). Environmental awareness was assessed via a concurrent object detection task embedded within the VR environment.
Analyses of the experimental flight demonstrated significant group differences in environmental awareness. Participants in the stress condition showed reduced object detection accuracy compared to controls (B= −0.39, SE= 0.16, z= −2.37, p= .018; OR= 0.68, 95% CI [0.50, 0.94]). Also, participants in the stress condition showed greater altitude deviation (β = 84.48 ft, SE = 41.14 ft, t(17) = 2.05, p = .056, R² = .20), though this effect did not reach statistical significance.

These findings suggest that acute stress impairs environmental awareness and may influence motor flight control in novice pilots. The study furthermore demonstrates the feasibility of integrating validated stress induction into VR flight simulation, offering an ecologically relevant platform for investigating stress-performance dynamics in aviation.
ORAL

During flight maneuvers, pilots are required to continuously sample and integrate information from multiple cockpit instruments in order to maintain accurate control of the aircraft's trajectory (Endsley, 1995). This monitoring places high demands on attention and cognitive resources, yet it is essential for quickly detecting and correcting deviations in key flight parameters. Recent technological advances have made eye-tracking easier to deploy in training and operational settings, providing a practical way to better understand how pilots allocate visual attention while flying (Knabl-Schmitz et al., 2023). Although many studies have examined visual scanning in commercial aviation simulators (Lefrançois et al., 2016; Haslbeck & Zhang, 2017; Lounis et al., 2021), our work extends this research by studying pilots in a simulator used for French Air Force training, allowing an in-context investigation within a military environment (Vantrepotte et al., 2025). We compare pilots with different levels of expertise (novices, cadets, and instructors) across a broader set of maneuvers, including turns, descents, climbs, and level flight.
A key contribution of this paper is that it goes beyond standard eye-tracking indicators (such as fixation duration or fixation count) and instead uses more advanced measures that can better capture how monitoring is organised. In practice, we assess not only how much pilots look at different instruments, but also how their visual attention is distributed (K coefficient and entropy) and how structured their visual strategies (Lempel–Ziv complexity and N-gram patterns) are over time, then relate these patterns to flight performance using deviation measures derived from flight parameters. We expect expertise to be associated with more efficient and more consistent monitoring strategies, and we argue that these advanced measures are particularly useful for detecting subtle differences between groups that may appear similar when using only basic metrics. Ultimately, this approach aims to refine our understanding of expertise-related monitoring strategies and support the design of training interventions better tailored to pilots at different stages of learning.

In the 1970s–early 1980s, Swedish military flight training underwent a significant pedagogical transformation. At the time, the system relied on hierarchical instruction and an attrition-based training philosophy. Over time, this approach became associated with a rising number of incidents and concerns regarding trust, instructional quality and flight safety. In response, the Swedish Air Force initiated a large-scale reform of both flight safety practices and organisational culture. This long-term effort led to a shift towards a reflective, dialogue-oriented and trust-based training culture that continues to shape Swedish military flight training.
The transformation included reforms in pilot selection and the professionalisation of flight instructors including a broader shift towards mission command and individual responsibility. These changes formed a pedagogical tradition that has gained international recognition in high-stakes military aviation contexts.
This historical analytical case study describes and analyses the development of Swedish military aviation pedagogy from the 1970s to the present. It explores its relevance for contemporary and emerging training environments. Including the Swedish jet fighter Saab JAS 39 Gripen, increasingly automated cockpit systems, NATO integration and Live Virtual Constructive (LVC) training concepts. LVC training integrates live aircraft operations with simulators and computer-generated entities to create integrated training environments. The analysis draws on perspectives from reflective learning and learning-organisation theory as well as research on team dynamics and organisational trust.
Preliminary findings indicate that the Swedish approach has shown strong continuity over time. It is characterised by structured debriefing, transparent communication, psychologically safe instructor–student relationships and the systematic development of shared mental models. These practices have historically supported both aviation safety and operational adaptability. As training environments become more complex and cognitively demanding, particularly within LVC contexts, the findings suggest that reflective and relational approaches are not diminished by technological development but instead become increasingly important. The study discusses implications for future military jet fighter training.

The brain is capable of high cognitive performance only when periods of effort are regularly interrupted by adequate recovery. This is particularly relevant in safety-critical domains such as air traffic control (ATC). Air traffic controllers (ATCOs) are responsible for the safe and orderly movement of aircraft in controlled airspace, a task that requires continuous situation monitoring, quick situation assessment, decision-making, and conflict resolution. To support cognitive performance, time on position is operationally limited. During a shift, ATCOs work in defined periods of up to 90 minutes on position, which are regularly followed by scheduled recovery breaks of roughly 60 minutes. As a result, breaks account for roughly 40 percent of an average operational shift. Despite this substantial proportion of off-position time during a shift, little is known about how ATCOs use these breaks and how specific break activities contribute to recovery.

The present study examined how characteristics of break activities relate to recovery experiences. Data were collected through repeated online surveys among Scandinavian ATCOs following operational early and late shifts. The surveys captured break activities, recovery experiences, subjective sleepiness, and need for recovery. Break activities were assessed across nine dimensions (e.g., physical, social, virtual, outdoor, relaxation-oriented, work-related). Recovery experiences were measured using four dimensions from the DRAMMA framework (Newman, Tay, & Diener, 2014): psychological detachment, relaxation, autonomy, and mastery. Post shift outcomes included sleepiness and need for recovery.

By linking break activities to psychological recovery processes post shift, the study aims to clarify how break activities contribute to effective recovery in operational ATC and whether these relationships differ between shift types. The findings contribute to a better understanding of at work recovery in ATC operations.

Low workload is a common problem in air traffic control (ATC) operations, potentially decreasing air traffic controllers' (ATCOs) level of vigilance, and their ability to complete operational tasks in a safe and efficient manner due to the lack of engagement. As both subjective questionnaires and physiological measurements such as pupillometry can capture underload, this study aimed to investigate underload in radar-based ATC during a calm rostering schedule in a European ATSP, to answer the following research questions: How do traffic and rostering parameters influence perceived underload, how is subjective underload reflected in pupillometry, and to what extent can pupillometry detect changes in ATCO engagement under low workload conditions that are not reflected in subjective workload ratings?

As such, ATCOs (N = 22) from four different radar-based working positions (e.g. Area, VFR) took part in a two-week measurement campaign involving both subjective (questionnaires) and objective (eye tracking) measures. ATCOs were asked to provide their sleep-wake and duty schedules every morning. In case of an operational duty, participants were also asked to rate their average and peak workload and alertness hourly during their duty, and traffic complexity and intensity for the shift as a whole. To validate the subjective workload and fatigue inputs, eye tracking was used during specific duties, using eye-related parameters such as blinks, pupil diameter and eyelid opening.

The results of the first measurement campaign showed that of the 116 recorded days, 72 days included an operational duty. However, for only 42 duties, data on traffic intensity and complexity was reported by participants. Sustained underload was reported on 33 of these shifts, with an average workload score of 34.56 (SD = 30.40) and peak workload at 48.30 (SD = 36.75) on the Rating Scale Mental Effort (RSME, ranging from 0 to 150) across all operational shifts, confirming underload. Analyses on the eye tracking data and traffic parameters are still ongoing, and results will be presented at the conference together with the study's implications.

Air traffic controllers (ATCOs) perform demanding 24/7 tasks where fatigue, though regulated, still impairs vigilance, attention, and decision‑making (Hudson et al., 2019; Killgore et al., 2006; Pauletti et al., 2024). Fatigue‑aware rostering initiatives, including EUROCONTROL guidance (2023), link roster design to fatigue risk, yet biomathematical models assume uniform off-duty sleep. Empirical evidence shows that controllers' sleep timing varies across shifts, underscoring the need for real‑world sleep–wake data. Sleep diaries and actigraphy capture these variations and predict performance fluctuations (Gradisar et al., 2007; Harris et al., 2021), reinforcing the importance of sleep timing and regularity beyond duration.
Sleep probability functions (24h, 15‑min resolution) were derived from sleep diaries of 16 ATCOs across multiple shift types (254 nights). Point‑wise k‑means functional clustering (k=5; R²=.46; Calinski–Harabasz=53.96) identified five sleep probability profiles: (1)bi‑phasic (n=27), (2)extended (n=45), (3)advanced (n=49), (4)regular (n=110), and (5)delayed (n=23).
Clusters and shift types were significantly associated (χ²(20)=205.54, p<.001). Night shifts mainly showed bi-phasic patterns (85.2%), rest days extended patterns (68.9%), morning shifts advanced ones (38.8%), and day shifts and rest days regular patterns (37.3% and 34.5%). The delayed cluster was primarily linked to night shifts (43.5%) and rest days (39.1%).
Subjective evaluations differed across both shifts and clusters. Morning and night shifts showed 1.2–1.8 points lower sleep quality and 1.1–1.9 points poorer daytime condition (all p<.001). Similarly, bi-phasic, advanced, and delayed patterns showed 1.3–2.3 points lower sleep quality and 1.2–2.2 points poorer daytime condition (all p<.001). Subjective sleepiness (Karolinska Sleepiness Scale) was significantly lower in the extended and regular clusters than in bi-phasic, advanced, and delayed clusters.
Linear Mixed Model analyses on a 3-min Behavioral Sleep Resistance Task (Mairesse et al., 2009) showed cluster-dependent differences in neurobehavioral performance. Advanced patterns exhibited greater reaction-time (RT) variability and interpercentile range, bi-phasic and advanced patterns showed higher lapse frequency (RT>355ms) and regular patterns showed more responses within the optimal performance domain (all p<.05).
Overall, regular patterns were associated with lower subjective sleepiness and better sustained attention, whereas advanced patterns showed reduced neurobehavioral stability and bi-phasic patterns increased attentional lapses. These profiles closely mirror ATCOs shift schedules and may inform fatigue‑risk management.

Artificial Intelligence is increasingly embedded in commercial flight operations through adaptive automation and decision-support systems. In response, the EU Artificial Intelligence Act and EASA's AI Roadmap 2.0 classify aviation AI applications as high-risk systems requiring demonstrable trustworthiness, human oversight, meaningful human control, and operational reliability. However, empirical evidence remains limited regarding how pilots regulate reliance when algorithmic recommendations influence time-critical decisions.
This study proposes a controlled experimental investigation examining how varying levels of AI transparency affect reliance behavior, cognitive workload, and intervention performance in high-fidelity flight simulation. The theoretical framework integrates Lee and See's trust in automation model with Parasuraman's levels of automation theory, conceptualizing reliance as a dynamic process shaped by perceived capability, explainability, and authority distribution.
The experimental manipulation systematically varies the informational architecture of AI-generated outputs across three interface conditions: directive recommendation without explanation; recommendation with probabilistic confidence indicators; and recommendation with structured rationale and alternatives. This design operationalizes transparency and explainability as measurable psychological variables.
A mixed factorial design is proposed in which transparency condition functions as a within-subject factor and pilot experience as a between-subject variable. An a priori power analysis (f = .25; α = .05) indicates that N = 72 pilots are required to achieve statistical power ≥ .80 in repeated-measures analyses. Scenarios are embedded within Line-Oriented Flight Training contexts including diversion planning, AI-supported fault diagnosis, and unstable approach management.
Dependent measures include adherence and override rates, intervention latency, decision quality, and physiological workload indices. Structured debriefings assess perceived authority and maintenance of meaningful human control.
The anticipated outcome is a Human–AI Oversight Performance Model identifying behavioral markers of appropriate reliance under algorithmic authority. The study provides empirical evidence directly applicable to Safety Management Systems, Evidence-Based Training, and AI certification processes in regulated aviation environments.

The integration of Artificial Intelligence (AI) into aviation marks a decisive turning point . Integrating AI into aviation offers significant potential to address these pressing challenges as Europe's shortage of air traffic controllers of 600–700 air traffic controllers, while traffic volumes continue to rise. However, the successful deployment of AI in high-risk domains such as air traffic control demands rigorous adherence to safety, security, human factors, and ethical principles—particularly ensuring robust human oversight.

The European Union Aviation Safety Agency (EASA) plays a leading role in shaping this transition within the aviation domain, developing guidelines aligned with the EU AI Act to ensure human-centered AI integration. As part of this, the EASA conducted a survey with aviation professionals (n=231) on ethics for AI. understanding the level of comfort, trust and acceptance of AI-based systems applied to the different aviation domains.

In parallel, the DLR research project LOKI (Human-AI Collaboration in Aviation) investigates and develops prototypes and guidelines for trustworthy human-AI collaboration in aviation. In this interdisciplinary effort, we are developing, in close collaboration with industry partners, an AI-based digital partner for human air traffic controllers. To ensure that user perspectives are embedded from the earliest design stages through to the final prototype, DLR involved air traffic controllers from Deutsche Flugsicherung GmbH (DFS) and Austro Control through four user workshops conducted between 2022 and 2026 (n=10), a user survey (n=170), and a validation study (n=10). At each step, the perspective of aviation operators on their future roles, the psycho-social requirements on humans and the framework conditions for human oversight were captured.

By synthesizing findings from the EASA ethics study involving aviation professionals with insights from the LOKI user studies, a comprehensive understanding of the expectations, concerns, as well as operational, psychological and social requirements for humans working with and overseeing AI-systems in future aviation is worked out. This integrated analysis provides critical guidance for the design, regulation, and implementation of trustworthy AI systems in air traffic management, paving the way for a safer, more resilient, and human-in-the-loop future of aviation.